a product of MVTec

Solution Guide II-D

Classification

HALCON 22.11 Steady

How to use classification, Version [Link]
All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means,
electronic, mechanical, photocopying, recording, or otherwise, without prior written permission of the publisher.

Copyright © 2008-2022 by MVTec Software GmbH, Munich, Germany MVTec Software GmbH

Protected by the following patents: US 7,239,929, US 7,751,625, US 7,953,290, US 7,953,291, US 8,260,059, US 8,379,014,
US 8,830,229, US 11,328,478. Further patents pending.
Microsoft, Windows, Windows 8.1, 10 (x64 editions), 11, Windows Server 2012 R2, 2016, 2019, 2022 Microsoft .NET, Visual
C++ and Visual Basic are either trademarks or registered trademarks of Microsoft Corporation.
All other nationally and internationally recognized trademarks and tradenames are hereby recognized.

More information about HALCON can be found at: [Link]

About This Manual

In a broad range of applications classification is suitable to find specific objects or detect defects in images. This
Solution Guide leads you through the variety of approaches that are provided by HALCON.
After a short introduction to the general topic in section 1 on page 7, a first example is presented in section 2 on
page 11 that gives an idea on how to apply a classification with HALCON.
Section 3 on page 15 then provides you with the basic theories related to the available approaches. Some hints how
to select the suitable classification approach, a set of features or images that is used to define the class boundaries,
and some samples that are used for the training of the classifier are given in section 4 on page 27.
Section 5 on page 31 describes how to generally apply a classification for various objects like pixels or regions
based on various features like color, texture, or region features. Section 6 on page 57 shows how to apply classi-
fication for a pure pixel-based image segmentation and section 7 on page 75 provides a short introduction to the
classification for optical character recognition (OCR). For the latter regions are classified by region features.
Finally, section 8 on page 93 provides some general tips that may be suitable when working with complex classi-
fication tasks.
The HDevelop example programs that are presented in this Solution Guide can be found in the specified subdirec-
tories of the directory %HALCONEXAMPLES%. The path to this directory can be determined with the operator call
get_system ('example_dir', ExampleDir).

Symbols

The following symbol is used within the manual:

! This symbol indicates an information you should pay attention to.

Contents

1 Introduction 7

2 A First Example 11

3 Classification: Theoretical Background 15

3.1 Classification in General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
3.2 Euclidean and Hyperbox Classifiers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
3.3 Multi-Layer Perceptrons (MLP) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
3.4 Support-Vector Machines (SVM) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
3.5 Gaussian Mixture Models (GMM) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
3.6 K-Nearest Neighbors (k-NN) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
3.7 Deep Learning (DL) and Convolutional Neural Networks (CNNs) . . . . . . . . . . . . . . . . . 23

4 Decisions to Make 27
4.1 Select a Suitable Classification Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
4.2 Select Suitable Features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
4.3 Select Suitable Training Samples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

5 Classification of General Features 31

5.1 General Approach (Classification of Arbitrary Features) . . . . . . . . . . . . . . . . . . . . . . . 31
5.2 Involved Operators (Overview) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
5.2.1 Basic Steps: MLP, SVM, GMM, and k-NN . . . . . . . . . . . . . . . . . . . . . . . . . 34
5.2.2 Advanced Steps: MLP, SVM, GMM, and k-NN . . . . . . . . . . . . . . . . . . . . . . . 35
5.3 Parameter Setting for MLP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
5.3.1 Adjusting create_class_mlp . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
5.3.2 Adjusting add_sample_class_mlp . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
5.3.3 Adjusting train_class_mlp . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
5.3.4 Adjusting evaluate_class_mlp . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
5.3.5 Adjusting classify_class_mlp . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
5.4 Parameter Setting for SVM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
5.4.1 Adjusting create_class_svm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
5.4.2 Adjusting add_sample_class_svm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
5.4.3 Adjusting train_class_svm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
5.4.4 Adjusting reduce_class_svm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
5.4.5 Adjusting classify_class_svm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
5.5 Parameter Setting for GMM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
5.5.1 Adjusting create_class_gmm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
5.5.2 Adjusting add_sample_class_gmm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
5.5.3 Adjusting train_class_gmm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
5.5.4 Adjusting evaluate_class_gmm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
5.5.5 Adjusting classify_class_gmm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
5.6 Parameter Setting for k-NN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
5.6.1 Adjusting create_class_knn . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
5.6.2 Adjusting add_sample_class_knn . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
5.6.3 Adjusting train_class_knn . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
5.6.4 Adjusting set_params_class_knn . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
5.6.5 Adjusting classify_class_knn . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
6 Classification for Image Segmentation 57
6.1 Approach for MLP, SVM, GMM, and k-NN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
6.1.1 General Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
6.1.2 Involved Operators (Overview) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
6.1.3 Parameter Setting for MLP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
6.1.4 Parameter Setting for SVM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
6.1.5 Parameter Setting for GMM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
6.1.6 Parameter Setting for k-NN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
6.1.7 Classification Based on Look-Up Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
6.2 Approach for a Two-Channel Image Segmentation . . . . . . . . . . . . . . . . . . . . . . . . . . 72
6.3 Approach for Euclidean and Hyperbox Classification . . . . . . . . . . . . . . . . . . . . . . . . 73

7 Classification for Optical Character Recognition (OCR) 75

7.1 General Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
7.2 Involved Operators (Overview) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
7.3 Parameter Setting for MLP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
7.3.1 Adjusting create_ocr_class_mlp . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
7.3.2 Adjusting write_ocr_trainf / append_ocr_trainf . . . . . . . . . . . . . . . . . . 82
7.3.3 Adjusting trainf_ocr_class_mlp . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
7.3.4 Adjusting do_ocr_multi_class_mlp . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
7.3.5 Adjusting do_ocr_single_class_mlp . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
7.4 Parameter Setting for SVM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
7.4.1 Adjusting create_ocr_class_svm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
7.4.2 Adjusting write_ocr_trainf / append_ocr_trainf . . . . . . . . . . . . . . . . . . 85
7.4.3 Adjusting trainf_ocr_class_svm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
7.4.4 Adjusting do_ocr_multi_class_svm . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
7.4.5 Adjusting do_ocr_single_class_svm . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
7.5 Parameter Setting for k-NN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
7.5.1 Adjusting create_ocr_class_knn . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
7.5.2 Adjusting write_ocr_trainf / append_ocr_trainf . . . . . . . . . . . . . . . . . . 87
7.5.3 Adjusting trainf_ocr_class_knn . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
7.5.4 Adjusting do_ocr_multi_class_knn . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
7.5.5 Adjusting do_ocr_single_class_knn . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
7.6 Parameter Setting for CNNs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
7.6.1 Adjusting do_ocr_multi_class_cnn . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
7.6.2 Adjusting do_ocr_single_class_cnn . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
7.7 OCR Features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89

8 General Tips 93
8.1 Optimize Critical Parameters with a Test Application . . . . . . . . . . . . . . . . . . . . . . . . 93
8.2 Classify General Regions using OCR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
8.3 Visualize the Feature Space (2D and 3D) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
8.3.1 Visualize the 2D Feature Space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
8.3.2 Visualize the 3D Feature Space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99

Index 105
 Introduction D-7

Introduction
Chapter 1

Introduction

What is Classification?
Classifying an object means to assign an object to one of several available classes. When working with images, the
objects usually are pixels or regions. Objects are described by features, which comprise, e.g., the color or texture
for pixel objects, and the size or specific shape features for region objects. To assign an object to a specific class,
the individual class boundaries have to be known. These are built in most cases by a training using the features
of sample objects for which the classes are known. Then, when classifying an unknown object, the class with
the largest correspondence between the feature values used for its training and the feature values of the unknown
object is returned.

What Can You Do With classification?

Classification is reasonable in all cases where objects have similarities, but within unknown variations. If you
search for objects of a certain fixed shape, and the points of a found contour may not deviate from this shape
more than a small defined distance, a template matching will be faster and easier to apply. But if the shapes
of your objects are similar, but you can not define exactly what the similarities are and what distinguishes these
objects from other objects in the image, you can show a classifier some samples of known objects (with a set of
features that you roughly imagine to describe the characteristics of the object types) and let the classifier find the
rules to distinguish between the object types. Classification can be used for a lot of different tasks. You can use
classification, e.g., for

• image segmentation, i.e., you segment images into regions of similar color or texture,

• object recognition, i.e., you find objects of a specific type within a set of different object types,
• quality control, i.e., you decide if objects are good or bad,
• novelty detection, i.e., you detect changes or defects of objects, or

• optical character recognition (OCR).

What can HALCON do for you?

To solve the different requirements on classification, HALCON provides different types of classifiers. The most
important HALCON classifiers are

• a classifier that uses neural nets, in particular multi-layer perceptrons (MLP, see section 3.3 on page 18),
• a classifier that is based on support-vector machines (SVM, see section 3.4 on page 19),
• a classifier that is based on Gaussian mixture models (GMM, see section 3.5 on page 20), and

• a classifier that is based on the k-nearest neighbors (k-NN, see section 3.6 on page 22).
• a classifier that is based on deep learning using a convolutional neural network (DL for general classification,
CNN for OCR, see section 3.7 on page 23).
D-8 Introduction

• Furthermore, for image segmentation also some simple but fast classifiers are available. These comprise a
classifier that segments two-channel images based on the corresponding 2D histogram (see section 6.2 on
page 72), a hyperbox classifier, and a classifier that can be applied using either a Euclidean or a hyperbox
metric (see section 3.2 on page 16 and section 6.3 on page 73).

For specific classification tasks, specific sets of HALCON operators are available. We distinguish between the
three following basic tasks:

• You can apply a general classification. Here, arbitrary objects like pixels or regions are classified based on
arbitrary features like color, texture, shape, or size. table 4.1 on page 28 may give a hint which method is
most suitable for your task. Section 5 on page 31 shows how to apply the suitable operators for MLP, SVM,
GMM, k-NN, and DL-based classification.
• You can apply classification for image segmentation. Here, the classification is used to segment images into
regions of different classes. For that, the individual pixels of an image are classified due to the features color
or texture and all pixels belonging to the same class are combined in a region. Section 6 on page 57 shows
how to apply the suitable operators for MLP, SVM, GMM, and k-NN classification (section 6.1 on page 57)
as well as for some simple but fast classifiers that segment the images using the 2D histogram of two image
channels (section 6.2 on page 72) or that apply an Euclidean or hyperbox classification (section 6.3 on page
73).

• You can apply classification for OCR, i.e., individual regions are investigated with regard to region features
and assigned to classes that typically (but not necessarily) represent individual characters or numbers. Sec-
tion 7 on page 75 shows how to apply the suitable operators for MLP, SVM, k-NN, and CNN classification.

What Are the Basic Steps of a Classification With HALCON?

There are different methods for classification implemented in HALCON, each one having its own assets and
drawbacks. For a brief comparison we refer to table 4.1 on page 28. These classification approaches can be
divided into two major groups. The first group consists of the methods MLP, SVM, GMM, and k-NN, where the
distinguishing features of each class have to be specified. The second group is given by DL-based methods, where
the network is trained by considering the inputs and outputs directly. For the user, it has the nice outcome of no
need for feature specification. Accordingly the basic approach for a classification with HALCON depends on the
method group. For the first one, thus MLP, SVM, GMM, and k-NN, it is as follows:

1. First, some sample objects, i.e., objects of known classes, are investigated. That is, a set of characteristic
features is extracted from each sample object and stored in a so-called feature vector (explicitly by the user
or implicitly by a specific operator).

2. The feature vectors of many sample objects are used to train a classifier. With the training, the classifier
derives suitable boundaries between the classes.
3. Then, unknown objects, i.e., the objects to classify, are investigated with the help of the same set of features
that was already used for the training samples. This step leads to feature vectors for the unknown objects.

4. Finally, the trained classifier uses the class boundaries that were derived during the training to decide for the
new feature vectors to which classes they belong.

For deep-learning-based methods, to train the classifier (or rather the network) one does not have to specify the
features but to provide labeled (hence, already classified) images. Therefore the basic approach is as follows:

1. Providing data: For each class one wants the classifier to distinguish, much data in form of already labeled
images has to be provided.
2. Training: From this data the algorithm learns how to classify your images. This is achieved by retraining the
already pretrained, more general network. As a result, the network is adapted to your specific classification
task.
3. Inference phase: Classify images using the adapted network.
 D-9

What Information Do You Find in This Solution Guide?

This manual provides you with

Introduction
• basic theoretical background for the provided classifiers (section 3 on page 15),
• tips for the decision making, in particular tips for the selection of a suitable classification approach, the
selection of suitable training samples and, if needed, the selection of suitable features that describe the
objects to classify (section 4 on page 27),
• guidance for the practical application of classification for general classification (section 5 on page 31), image
segmentation (section 6 on page 57), and OCR (section 7 on page 75), and
• additional tips that may be useful when applying classification (section 8 on page 93). In particular, when
not applying deep-learning-based methods or only using it for OCR, tips how to adjust the most critical
parameters, tips how to use OCR for the classification of arbitrary regions, and tips how to visualize the
feature space for 2D and 3D feature vectors are provided.

What Do You Have to Consider Before Classifying?

Note that the decision which classifier to use for a specific application is a challenging task. There are no fixed
rules which approach works better for which application, as the number of possible fields of applications is very
large. At least, section 4.1 on page 27 provides some hints about the advantages and disadvantages of the individual
approaches.
Additionally, if you have decided to use a specific classifier, it is not guaranteed that you get a satisfying result
within a short time. Actually, in almost any case you have to apply a lot of tests with different parameters until you
get the result you aimed at. Classification is very complex! So, plan enough time for your application.
D-10 Introduction
 A First Example D-11

Chapter 2

First Example
A First Example

This section shows a first example for a classification that classifies metal parts based on selected shape fea-
tures. To follow the example actively, start the HDevelop program %HALCONEXAMPLES%\solution_guide\
classification\classify_metal_parts.hdev; the steps described below start after the initialization of the
application.
Step 1: Create classifier
First, a classifier is created. Here, we want to apply an MLP classification, so a classifier of type MLP is created
with create_class_mlp. The returned handle MLPHandle is needed for all following classification steps.

create_class_mlp (6, 5, 3, 'softmax', 'normalization', 3, 42, MLPHandle)

Step 2: Add training samples to the classifier

Then, the training images, i.e., images that contain objects of known class, are investigated. Each image contains
several metal parts that belong to the same class. The index of the class for a specific image is stored in the tuple
Classes. In this case, nine images are available (see figure 2.1). The objects in the first three images belong to
class 0, the objects of the next three images belong to class 1, and the last three images show objects of class 2.

FileNames := ['nuts_01', 'nuts_02', 'nuts_03', 'washers_01', 'washers_02', \

'washers_03', 'retainers_01', 'retainers_02', \
'retainers_03']
Classes := [0, 0, 0, 1, 1, 1, 2, 2, 2]

Now, each training image is processed by the two procedures segment and add_samples.

for J := 0 to |FileNames| - 1 by 1
read_image (Image, 'rings/' + FileNames[J])
segment (Image, Objects)
add_samples (Objects, MLPHandle, Classes[J])
endfor

The procedure segment segments and separates the objects that are contained in the image using a simple blob
analysis (for blob analysis see Solution Guide I, chapter 4 on page 35).

procedure segment (Image, Regions)

binary_threshold (Image, Region, 'max_separability', 'dark', UsedThreshold)
connection (Region, ConnectedRegions)
fill_up (ConnectedRegions, Regions)
return ()

For each region, the procedure add_samples determines a feature vector using the procedure get_features.
The feature vector and the known class index build the training sample, which is added to the classifier with the
operator add_sample_class_mlp.
D-12 A First Example

Class 0 Class 1 Class 2

Figure 2.1: Training images.

procedure add_samples (Regions, MLPHandle, Class)

count_obj (Regions, Number)
for J := 1 to Number by 1
select_obj (Regions, Region, J)
get_features (Region, Features)
add_sample_class_mlp (MLPHandle, Features, Class)
endfor
return ()

The features extracted in the procedure get_features are region features, in particular the ’circularity’,
’roundness’, and the four moments (obtained by the operator moments_region_central_invar) of the region.

procedure get_features (Region, Features)

select_obj (Region, SingleRegion, 1)
circularity (SingleRegion, Circularity)
roundness (SingleRegion, Distance, Sigma, Roundness, Sides)
moments_region_central_invar (SingleRegion, PSI1, PSI2, PSI3, PSI4)
Features := [Circularity,Roundness,PSI1,PSI2,PSI3,PSI4]
return ()

Step 3: Train the classifier

After adding all available samples, the classifier is trained with train_class_mlp.

train_class_mlp (MLPHandle, 200, 1, 0.01, Error, ErrorLog)

Step 4: Classify new objects

Now, images with different unknown objects are investigated. The segmentation of the objects and the extraction
of their feature vectors is realized by the same procedures that were used for the training images (segment and
get_features). But this time, the class of a feature vector is not yet known and has to be determined by the
classification. Thus, opposite to the procedure add_samples, within the procedure classify the extracted feature
vector is used as input to the operator classify_class_mlp and not to add_sample_class_mlp. The result is
the class index that is suited best for the feature vector extracted for the specific region.
 D-13

for J := 1 to 4 by 1
read_image (Image, 'rings/mixed_' + J$'02d')
segment (Image, Objects)
classify (Objects, MLPHandle, Classes)
disp_obj_class (Objects, Classes)
endfor

procedure classify (Regions, MLPHandle, Classes)

count_obj (Regions, Number)

First Example
Classes := []
for J := 1 to Number by 1
select_obj (Regions, Region, J)
get_features (Region, Features)
classify_class_mlp (MLPHandle, Features, 1, Class, Confidence)
Classes := [Classes,Class]
endfor
return ()

For a visual check of the result, the procedure disp_obj_class displays each region with a specific color that
depends on the class index (see figure 2.2).

Figure 2.2: Classifying metal parts because of their shape: (left) image with metal parts, (right) metal parts classified
into three classes (illustrated by different gray values).

procedure disp_obj_class (Regions, Classes)

count_obj (Regions, Number)
Colors := ['yellow', 'magenta', 'green']
for J := 1 to Number by 1
select_obj (Regions, Region, J)
dev_set_color (Colors[Classes[J - 1]])
dev_display (Region)
endfor
return ()
D-14 A First Example
 Classification: Theoretical Background D-15

Chapter 3

Classification: Theoretical
Background

Overview
This section introduces you to the basics of classification (section 3.1) and the specific classifiers that can be ap-
plied with HALCON. In particular, the Euclidean and hyperbox classifiers (section 3.2), the classifier based on
multi-layer perceptrons (neural nets, section 3.3), the classifier based on support-vector machines (section 3.4),
the classifier based on Gaussian mixture models (section 3.5), the classifier based on k-nearest neighbors (sec-
tion 3.6), and the classifier based on deep learning, with a focus on convolutional neural networks (section 3.7),
are introduced.

3.1 Classification in General

Generally, a classifier is used to assign an object to one of several available classes. For example, you have gray
value images containing citrus fruits. You have extracted regions1 from the images and each region represents
a fruit. Now, you want to separate the oranges from the lemons. To distinguish the fruits, you can apply a
classification. Then, the extracted regions of the fruits are your objects and the task of the classification is to decide
for each region if it belongs to the class ’oranges’ or to the class ’lemons’.
In order to decide to which class an image or a region belongs, the classifier needs to know how to distinguish the
classes. Thus, differences between the classes and the similarities within each individual class have to be known.
With deep-learning-based classification, the network learns this information automatically from the images. For
further information, see section 3.7 on page 23. However, for all other approaches, you as user need to provide the
knowledge, which you can obtain by analyzing typical features of the objects to classify. Let us illustrate the latter
case with the example of citrus fruits (an actual program is described in more detail in section 8.3.1 on page 96).
Suitable features can be, e.g., the ’area’ (an orange is usually bigger than a lemon) and the shape, in particular the
’circularity’ of the regions (the outline of an orange is closer to a circle than that of a lemon). Figure 3.1 shows
some oranges and lemons for which the regions are extracted and the region features ’area’ and ’circularity’
are calculated.
The features are arranged in an array that is called feature vector. The features of the feature vector span a so-
called feature space, i.e., a vector space in which each feature is represented by an axis. Generally, a feature space
can have any dimension, depending on the number of features contained in the feature vector. For visualization
purpose, here a 2D feature space is shown. In practice, feature spaces of higher dimension are very common.
In figure 3.2 the feature vectors of the fruits shown in figure 3.1 are visualized in a 2D graph, for which one axis
represents the ’area’ values and the other axis represents the ’circularity’ values. Although the regions vary
in size and circularity, we can see that they are similar enough to build clusters. The goal of a classifier is to
separate the clusters and to assign each feature vector to one of the clusters. Here, the oranges and lemons can
be separated, e.g., by a straight line. All objects on the lower left side of the line are classified as lemons and all
objects on the upper right side of the line are classified as oranges.
As we can see, the feature vector of a very small orange and that of a rather circular lemon are close to the
separating line. With a little bit different data, e.g., if the small orange additionally would be less circular, the
1 How to extract regions from images is described, e.g., in Solution Guide I, chapter 4 on page 35
D-16 Classification: Theoretical Background

Figure 3.1: Region features of oranges and lemons are extracted and can be added as samples to the classifier.

feature vectors may be classified incorrectly. To minimize errors, a lot of different samples and in many cases also
additional features are needed. An additional feature for the citrus fruits may be, e.g., the gray value. Then, not a
line but a plane is needed to separate the clusters. If color images are available, you can combine the area and the
circularity with the gray values of three channels. For feature vectors of more than three features, an n-dimensional
plane, also called hyperplane, is needed.
Classifiers that use separating lines or hyperplanes are called linear classifiers. Other classifiers, i.e., non-linear
classifiers, can separate clusters using arbitrary surfaces and may be able to separate clusters more conveniently in
some cases.
Summarized, we need a suitable set of features and we have to select the classifier that is suited best for a spe-
cific classification application. To select the most appropriate approach, we have to know some basics about the
available classifiers and the algorithms they use.

3.2 Euclidean and Hyperbox Classifiers

One of the simple classifiers is the Euclidean or minimum distance classifier. With HALCON, the Euclidean
classification is available for image segmentation, i.e., the objects to classify are pixels and the feature vectors
contain the gray values of the pixels. The dimension of the feature space depends on the number of channels used
for the image segmentation. Geometrically interpreted, this classifier builds circles (in 2D; see figure 3.3a) or n-
dimensional hyperspheres (in nD) around the cluster centers to separate the clusters from each other. In section 6.3
on page 73 it is described how to apply the Euclidean classifier for image segmentation. With HALCON, the
Euclidean metric is used only for image segmentation, not for the classification of general features or OCR. This
is because the approach is stable only for feature vectors of low dimension.
Whereas the Euclidean classifier uses n-dimensional spheres, the hyperbox approach uses axis-parallel cubes, so-
called hyperboxes (see figure 3.3b). This can be imagined as a threshold approach in multidimensional space.
 3.2 Euclidean and Hyperbox Classifiers D-17

Overview
1

Figure 3.2: The normalized values for the ’area’ and ’circularity’ of the fruits span a feature space. The two
classes can be separated by a line.

Feature 1: Feature 1:
x x
x xx xx x xx xx
xx x x xx x x x
x xxx x xx
x x xx x x x xx x
xxx xxx

x x x xx x x x xx x x x
x x x xx x x x xxx x
xx x xx xx x
x xxx xx x xxxxxx

Feature 2: Feature 2:

(a) (b)

Figure 3.3: (a) Euclidean classifier and (b) hyperbox classifier.

That is, for each class specific value ranges for each axis of the feature space are determined. If a feature vector
lies within all the ranges of a specific class, it will be assigned to this class. The hyperboxes can overlap. For
objects that are ambiguous, the hyperbox approach can be combined with another classification approach, e.g., an
Euclidean classification or a maximum likelihood classification. Within HALCON, the Euclidean distance is used
and additionally weighted with the variance of the feature vector. In section 6.3 on page 73 it is described how to
apply the hyperbox classifier for image segmentation.
HALCON provides also operators for hyperbox classification of general features as well as for OCR, but these
show almost no advantage but a lot of disadvantages compared to the MLP, SVM, GMM, and k-NN approaches,
and thus are not described further in this solution guide.
D-18 Classification: Theoretical Background

3.3 Multi-Layer Perceptrons (MLP)

Neural nets directly determine the separating hyperplanes between the classes. For two classes the hyperplane
actually separates the feature vectors of the two classes, i.e., the feature vectors that lie on one side of the plane are
assigned to class 1 and the feature vectors that lie on the other side of the plane are assigned to class 2. In contrast
to this, for more than two classes the planes are chosen such that the feature vectors of the correct class have the
largest positive distance of all feature vectors from the plane.
A linear classifier can be built, e.g., using a neural net with a single layer like shown in figure 3.4 (a,b). There,
so-called processing units (neurons) first compute the linear combinations of the feature vectors and the network
weights and then apply a nonlinear activation function.
A classification with single-layer neural nets needs linearly separable classes, which is not sufficient in many
classification applications. To get a classifier that can separate also classes that are not linearly separable, you can
add more layers, so-called hidden layers, to the net. The obtained multi-layer neural net (see figure 3.4, c) then
consists of an input layer, one or several hidden layers and an output layer. Note that one hidden layer is sufficient
to approximate any separating hypersurface and any output function with values in [0,1] as long as the hidden layer
has a sufficient number of processing units.

Single−layer neural networks Multi−layer neural network

x1 w1 x1 x1
x2 w2 x2 x2

wn b
xn xn
xn
Two−class single−layer neural net n−class single−layer neural net

a) b) c)

Figure 3.4: Neural networks: single-layered for (a) two classes and (b) n classes, (c) multi-layered: (from left to right)
input layer, hidden layer, output layer.

Within the neural net, the processing units of each layer (see figure 3.5) compute the linear combination of the
feature vector or of the results from a previous layer.

Input Hidden unit Output

weighted summation activation function

x1 w1
x2 w2
y

xn wn

Figure 3.5: Processing unit of an MLP.

That is, each processing unit first computes its activation as a linear combination of the input values:

nl
(l) (l) (l−1) (l)
X
aj = wij xi + bj
i=1

with
 3.4 Support-Vector Machines (SVM) D-19

• x0i : feature vector

(j)
• xi : result vector of layer l
(l) (l)
• wji and bj : weights of layer l

Then the results are passed through a nonlinear activation function:

(l) (l)
xj = f (aj )

With HALCON, for the hidden units the activation function is the hyperbolic tangent function:

ex − e−x
f (x) = tanh(x) =
ex + e−x

Overview
For the output function (when using the MLP for classification) the softmax activation function is used, which
maps the output values into the range (0, 1) such that they add up to 1:

exi
f (xi ) =
Σnj=1 exj

To derive the separating hypersurfaces for a classification using a multi-layer neural net, the network weights
have to be adjusted. This is done by a training. That is, data with known output is inserted to the input layer
and processed by the hidden units. The output is then compared to the expected output. If the output does not
correspond to the expected output (within a certain error tolerance), the weights are incrementally adjusted so that
the error is minimized. Note that the weight adjustment using HALCON is realized by a very stable numeric
algorithm that leads to better results than obtained by the classical back propagation algorithm.
The MLP method works for classification of general features, image segmentation, and OCR. Note that MLP can
also be used for least squares fitting (regression) and for classification problems with multiple independent logical
attributes.
An MLP can have more than one hidden layer and is then considered as a deep learning method. In HALCON
we only have a single hidden layer implemented in our MLPs. That is why, whenever we refer to deep learning
methods, we exclude the MLP method.

3.4 Support-Vector Machines (SVM)

Another classification approach that can handle classes that are not linearly separable uses support-vector machines
(SVM). Here, no non-linear hypersurface is obtained, but the feature space is transformed into a space of higher
dimension, so that the features become linearly separable. Then, the feature vectors can be classified with a linear
classifier.
In figure 3.6, e.g., two classes in a 2D feature space are illustrated by black and white squares, respectively. In the
2D feature space, no line can be found that separates the classes. When adding a third dimension by deforming the
plane built by Feature1 and Feature2, the classes become separable by a plane.
To avoid the curse of dimensionality (see section 3.5) for SVM, not the features but a kernel is transformed. The
challenging task is to find the suitable kernel to transform the feature space into a higher dimension so that the
black squares in figure 3.6 go up and the white ones stay in their place (or at least stay in another value range of
the axis for the additional dimension). Common kernels are, e.g., the inhomogeneous polynomial kernel or the
Gaussian radial basis function kernel.
With SVM, the separating hypersurface for two classes is constructed such that the margin between the two classes
becomes as large as possible. The margin is defined as the closest distance between the separating hyperplane and
any training sample. That is, several possible separating hypersurfaces are tested and the surface with the largest
margin is selected. The training samples from both classes that have exactly the closest distance to the hypersurface
are called ’support vectors’ (see figure 3.7 for two linearly separable classes).
D-20 Classification: Theoretical Background

Feature 1: Additional Dimension

Feature 1

Separating Hyperplane

Feature 2: Feature 2

Figure 3.6: Separate two classes (black and white squares): (left) In the 2D feature space the classes can not
be separated by a straight line, (right) by addition of a further dimension, the classes become linearly
separable.

Support vectors

Hyperplane

Figure 3.7: Support vectors are those feature vectors that have exactly the closest distance to the hyperplane.

By nature SVM can handle only two-class problems. Two approaches can be used to extend the SVM to a multi-
class problem: With the first approach pairs of classes are built and for each pair a binary classifier is created.
Then, the class that wins most of the comparisons is the best suited class. With the second approach, each class
is compared to the rest of the training data and then, the class with the maximum distance to the hypersurface is
selected (see also section 5.4.1 on page 44).
SVM works for classification of general features, image segmentation, and OCR.

3.5 Gaussian Mixture Models (GMM)

The theory for the classification with Gaussian mixture models (GMM) is a bit more complex. One of the basic
theories when dealing with classification comprises the Bayes decision rule. Generally, the Bayes decision rule
tells us to minimize the probability of erroneously classifying a feature vector by maximizing the probability for
the feature vector x to belong to a class. This so-called ’a posteriori probability’ should be maximized over all
classes. Then, the Bayes decision rule partitions the feature space into mutually disjoint regions. The regions are
separated by hypersurfaces, e.g., by points for 1D data or by curves for 2D data. In particular, the hypersurfaces
are defined by the points in which two neighboring classes are equally probable.
The Bayes decision rule can be expressed by

P (x|wi ) × P (wi )
P (wi |x) =
P (x)

with

• P (wi |x): a posteriori probability

 3.5 Gaussian Mixture Models (GMM) D-21

• P (x|wi ): a priori probability that the feature vector x occurs given that the class of the feature vector is wi
• P (wi ): Probability, that the class wi occurs

• P (x): Probability that the feature vector x occurs

For classification, the a posteriori probability should be maximized over all classes. Here, we coarsely show how
to obtain the a posteriori probability for a feature vector x. First, we can remark that P (x), i.e., the probability of
the class, is a constant if x exists.
The first problem of the Bayes classifier is how to obtain P (wi ), i.e., the probability of the occurrence of a class.
Two strategies can be followed. First, you can estimate it from the used training set. This is recommended only if
you have a training set that is representative not only with regard to the quality of the samples but also with regard
to the frequency of the individual classes inside the set of samples. As this strategy is rather uncertain, a second
strategy is recommended in most cases. There, it is assumed that each class has the same probability to occur, i.e.,
P (wi ) is set to 1/m with m being the number of available classes.
The second problem of the Bayes classifier is how to obtain the a priori probability P (x|wi ). In principle, a his-

Overview
togram over all feature vectors of the training set can be used. The apparent solution is to subdivide each dimension
of the feature space into a number of bins. But as the number of bins grows exponentially with the dimension of the
feature space, you face the so-called ’curse of dimensionality’. That is, to get a good approximation for P (x|wi ),
you need more memory than can be handled properly. With another solution, instead of keeping the size of a bin
constant and varying the number of samples in the bin, the number of samples k for a class wi is kept constant
while varying the volume of the region in space around the feature vector x that contains the k samples (v(x, wi )).
The volume depends on the k nearest neighbors of the class wi , so the solution is called k nearest-neighbor density
estimation. It has the disadvantage that all training samples have to be stored with the classifier and the search for
the k nearest neighbors is rather time-consuming. Because of that, it is seldom used in practice. A solution that
can be used in practice assumes that P (x|wi ) follows a certain distribution, e.g., a normal distribution. Then, you
only have to estimate the two parameters of the normal distribution, i.e., the mean vector µi and the covariance
matrix Σi . This can be achieved, e.g., by a maximum likelihood estimator.
In some cases, a single normal distribution is not sufficient, as there are large variations inside a class. The character
’a’, e.g., can be represented by ’a’ or ’a’, which have significantly different shapes. Nevertheless, both belong to
the same character, i.e., to the same class. Inside a class with large variations, a mixture of li different densities
exists. If these are again assumed to be normal distributed, we have a Gaussian mixture model. Classifying with
a Gaussian mixture model means to estimate to which specific mixture density a sample belongs. This is done by
the so-called expectation minimization algorithm.
Coarsely spoken, the GMM classifier uses probability density functions of the individual classes and expresses
them as linear combinations of Gaussian distributions (see figure 3.8). Comparing the approach to the simple clas-
sification approaches described in section 3.2 on page 16, you can imagine the GMM to construct n-dimensional
error (covariance) ellipsoids around the cluster centers (see figure 3.9).

Class 1 Class 2

Feature Vectors
Feature Vector X

Figure 3.8: The variance of class 1 is significantly larger than that of class 2. In such a case, the distance to the
Gauss error distribution curve is a better criteria for the class membership than the distance to the
cluster center.

GMM are reliable only for low dimensional feature vectors (approximately up to 15 features), so HALCON pro-
vides GMM only for the classification of general features and image segmentation, but not for OCR. Typical
Applications are image segmentation and novelty detection. Novelty detection is specific for GMM and means
that feature vectors that do not belong to one of the trained classes can be rejected. Note that novelty detection can
D-22 Classification: Theoretical Background

Feature 1

Feature Vector X

Class 2

Class 1
Feature 2

Figure 3.9: The feature vector X is nearer to the error ellipse of class 1 although the distance to the cluster center of
class 1 is larger than the distance to the cluster center of class 2.

also be applied with SVM, but then a specific parameter has to be set and only two-class problems can be handled,
i.e., a single class can be trained and the feature vectors that do not belong to that single class are rejected.
There are two general approaches for the construction of a classifier. First, you can estimate the a posteriori
probability from the a priori probabilities of the different classes (statistical approach), which we have introduced
here for classification with the GMM classifier. Second, you can explicitly construct the separating hypersurfaces
between the classes (geometrical approach). This can be realized in HALCON either with a neural net using
multi-layer perceptrons (see section 3.3 on page 18) or with support-vector machines (see section 3.4 on page 19).

3.6 K-Nearest Neighbors (k-NN)

K-Nearest Neighbors (k-NN) is a simple yet powerful approach that stores the features and classes of all given
training data and classifies each new sample based on its k-nearest neighbors in the training data.
The following example illustrates the basic principle of k-NN classification. Here, a two dimensional feature space
is used, i.e., each training sample consists of two feature values and a class label (see figure 3.10). The two classes
A and B are represented by three training samples, each. We can now use the training data to classify the new
sample N. For this, the k-nearest neighbors of N are determined in the training data.

Feature 1

B2
B1
A1

N
A2
B3

1 A3
Feature 2
1

Figure 3.10: Example for k-NN classification. Class A is represented by the three samples A1 , A2 , and A3 , and
class B is represented by the three samples B1 , B2 , and B3 . The class of the new sample N is to be
determined with k-NN classification.

If we are using k=1, only the nearest neighbor of N is determined and we can directly assign its class label to the
new sample. Here, the training sample A2 is closest to N. Therefore, the new sample N is classified as being of
class A.
 3.7 Deep Learning (DL) and Convolutional Neural Networks (CNNs) D-23

In case k is set to a value larger than 1, the class of the new sample N must be derived from its k-nearest neighbors
in the training data. The two approaches, which are most frequently used for this task, are a simple majority vote
and a weighted majority vote that takes into account the distances to the k nearest neighbors.
For example, if we are using k=3, we need to determine the three nearest neighbors of N. In the above example,
the distances from N to the training samples are:

Distance
A1 5.2
A2 1.1
A3 4.7
B1 2.8
B2 4.2
B3 3.1

Thus, the three nearest neighbors of N are A2 , B1 , and B3 .

Overview
A simple majority vote would assign class B to the new sample N, because two of the three nearest neighbors of N
belong to the class B.
The weighted majority vote takes into account the distances from N to the k-nearest neighbors. In the example,
class A would be assigned to N, because N lies very close to A2 and significantly further away from B1 and B3 .
Despite the simplicity of this approach, k-NN typically yields very good classification results. One big advantage
of the k-NN classifier is that it works directly on the training data, which leads to a blazingly fast training step. Due
to this, it is especially well suited for testing various configurations of training data. Furthermore, newly available
training data can be added to the classifier at any time. However, the classification itself is slower than, e.g., the
MLP classification, and the k-NN classifier may consume a lot of memory because it contains the complete training
data.

3.7 Deep Learning (DL) and Convolutional Neural Networks

(CNNs)
The term "deep learning" was originally used to describe the training of neural networks with multiple hidden lay-
ers. Today it is rather used as a generic term for several different concepts in machine learning. Only recently, with
the advent of processing power, large datasets, and proper algorithms, it led to breakthroughs in many applications.
One particular successful example is image classification based on CNNs (Convolutional Neural Networks), char-
acterized by the presence of at least one convolutional layer in the network. CNNs are inspired by the visual cortex
of humans and animals. When we see edges with certain orientations, some individual neural cells in the brain
respond. Some neurons, e.g., fire when exposed to vertical edges and some when shown horizontal or diagonal
edges. Similarly, convolutional neural networks perform classification by looking for low level features, like edges
and curves, and then building up to more abstract concepts. These concepts might be similar to text, logos, or
machine components. These features are selected automatically during the training.
As this Solution Guide is about classification, we will restrict this chapter to the deep learning method classification.
Note that there are also other deep learning methods, geared to fields of application according to their peculiarities.
For an overview of the different methods implemented in HALCON, please see the chapter “Deep Learning” in
the Reference Manual.
As already mentioned, CNNs used for deep-learning-based methods have multiple layers. A layer is a building
block performing specific tasks (e.g., convolution, pooling, etc., see below). It can be seen as a container, which
receives an input, applies an operation on it, and returns an output, for most layers feature maps. This output
serves as input for the next layer. Input and output layers are connected to the dataset, i.e., the image pixels or
the labels, respectively. The layers in between are called hidden layers. All layers together form a network, a
function mapping the input data onto classes. An illustration is shown in figure 3.11. Many of these layers, also
called filters, have weights, the filter weights. These are the parameters optimized when training a network. But
there are other, additional parameters, which are not directly learned during the regular training. These parameters
have values set before starting the training. We refer to this last type of parameters as hyperparameters in order to
distinguish them from the network parameters that are optimized during training. Note, training a network is not
a pure optimization problem: Machine learning usually acts indirectly. This means, we do not directly optimize
D-24 Classification: Theoretical Background

the mapping function predicting the classes. Instead, the loss function is introduced, a function penalizing the
deviation between the predicted and true classes. The loss function is now optimized, in the hope of doing so will
also improve our performance measure. Thus, training the network for the specific classification tasks, one strives
to minimize the loss (an error function) of the mapping function. In practice, this optimization is done calculating
the gradient and updating the parameters (weights) accordingly and iterating multiple times over the training data.
For more details we refer to the Reference Manual entry of the operator train_dl_model_batch.

Input Output

apple
lemon
orange

Feature map Feature maps Feature maps

Input Layer Hidden Layers Output Layer

Figure 3.11: A neural network as used for deep learning consists of multiple layers, potentially a huge number which
led to the name ’deep’ learning. The illustrated network classifies images (taken as input) by assigning
it a confidence value for each of the three distinguished classes (the output).

The network is trained by only considering the input and output, which is also called end-to-end learning. Ba-
sically, using the provided labeled images, the training algorithm adjusts the CNN filter weights such that the
network is able to distinguish the classes properly. For the user, it has the nice outcome of no need for manual
feature specification. Instead, however, the order, type, and number of layers of the neural network (also called
architecture of the network), as well as hyperparameters have to be specified. On top of this, for general classifica-
tion tasks, a lot of appropriate data has to be provided.
Currently, deep-learning-based classification can be used for two tasks within HALCON: a) for general classi-
fication, and b) for dedicated OCR classification. This differentiation is also reflected in the operator names,
where operators for general classification are part of the deep learning model and as a consequence marked with
dl_model while operators for OCR classification are marked with cnn.
Additionally, in the general case one can neither create a network from scratch nor create its own network archi-
tecture. Instead we use a technique called transfer learning, as will be explained below. In the OCR case it can
only be applied using the pretrained font Universal (see Solution Guide I,section 18.8 on page 222). That is, it is
not yet possible to train your own deep-learning-based OCR classifiers. In the following, some basic ideas on the
theory of CNN classifiers are described.
Building up and training a network from scratch takes a lot of time, computing power, expert knowledge, and a
huge amount of data. HALCON provides you with a trained network and uses a technique called transfer learning.
This means, we use a pretrained network, where the output layer is adapted to the respective application. Now, the
hidden layers are retrained for a specific task with potentially completely different classes. Thus, using transfer
learning, you will need fewer images and resources. More information about this can be found in the chapter
“Deep Learning” of the Reference Manual.
In the last stage, the inference phase, the network (which is now trained for your specific task) is applied to infer
input images. Unlike during training phase, the network is not changed anymore.
The classifier takes an image as input. But as an output it will not directly tell, that it belongs to a certain class.
Instead the classifier returns the inferred confidence values, expressing how likely the image belongs to every
distinguished class. E.g., the two classes ’apple’, and ’lemon’ are distinguished. Now we give an image of an
apple to the classifier. As a result, we get a confidence value for each class, like ’apple’: 0.97, and ’lemon’: 0.03.
To give you a basic idea about such a network, some common types of the hidden layers are introduced, in particular
convolutional, pooling, ReLU, and fully connected layers.
Convolutional layer
The first hidden layer is often a convolutional layer. Its functioning in a nutshell: A filter, also called kernel, is
moved across a feature map out of an other layer (which can be regarded as image and thus is sometimes named
as such), see figure 3.12. The covered part of the input feature map is taken, an operation applied, and the result
 3.7 Deep Learning (DL) and Convolutional Neural Networks (CNNs) D-25

determines the value of the corresponding output feature map entry. The kernel moves forward to select the next
part of the input feature map. Thereby, the stride determines how the kernel is moved, usually how many pixels
to the right and, once the end of the feature map width is reached, how many pixels down the next row is started.
The kernel itself is an array of given size, filled with numbers. These numbers are the filter’s weights, which are
learned during training.

Input feature map Output feature map

Figure 3.12: A 3x3 kernel is moved across a 6x6 feature map. The first selected feature map section is on the top
left corner, the second one with a stride of 1,1 on the top, but starting at the second pixel from the left.
The result is a 4x4 output feature map.

Overview
In convolutional layers the operation performed is a Hadamard-product: The pixel values of the feature map section
are multiplied element-wise with the filter weights and summed up. An example for the first two selected parts is
shown in figure 3.13. As this operation represents a convolution, the name of this layer is ’convolutional layer’.

Input feature map

0 0 0 -1 -1 -1 0*(-1)+0*(-1)+0*(-1)
0 2 3
4 9 9 * 0 0 0
1 1 1
+ 0*0 +2*0 +3*0
+ 4*1 +9*1 +9*1
=22

0 0 6 -1 -1 -1 0*(-1)+0*(-1)+6*(-1)
2 3 3
9 9 9 * 0 0 0
1 1 1
+ 2*0 +3*0 +3*0
+ 9*1 +9*1 +9*1
=21

Selected parts Filter

with pixel values with weights

22 21

Output feature map

Figure 3.13: A 3x3 kernel is moved across a 6x6 feature map with a 1,1 stride. Here, the calculation of the first two
entries is demonstrated. The shown filter is designed to look for horizontal edges, as such give rise to
a larger absolute values.

This convolution is performed for the whole input feature map. In figure 3.13, the filtering leads to a feature
map that provides information where to find horizontal edges in the input feature map. In practice, many dif-
ferent learned filters are used to determine features of the image. As a result, the individual filters produce two-
dimensional activation maps, which are then stacked along the third dimension to produce the three-dimensional
output volume (see figure 3.15). For more details we refer to the HALCON Reference Manual entry of the operator
create_dl_layer_convolution.
Pooling layer
D-26 Classification: Theoretical Background

Pooling is a form of non-linear down-sampling. It reduces the spatial size of the representation as well as the
number of parameters in the network and therefore the risk of overfitting. From the input feature map, which can
be regarded as image, a part with size of the kernel is taken. From this part, the maximum (’max pooling’) or
average (’average pooling’) is determined and put in the resulting feature map. The kernel ’moves’ as determined
by the stride and repeats the operation on the next part of the input feature map. Figure 3.14 illustrates two
examples with different pooling type and stride. As visible in the illustration, the resulting feature map has a size
depending on the input feature map, the kernel size, and the stride (and some further parts, e.g., padding). It is
possible to have the resulting feature map independent of the input feature map size, but in this case the kernel or
stride have to be adapted. Doing so is called ’global max pooling’ and ’global average pooling’, respectively. For
more details we refer to the HALCON Reference Manual entry of the operator create_dl_layer_pooling.

1 6 1 5 1 6 1 5
5 3.5 2
8 5 2 0 8 5 8 5 2 0
5 4 2
2 5 4 2 5 7 2 5 4 2
3 4.5 3.5
3 2 7 1 3 2 7 1

(1) (2)

Figure 3.14: Two examples for pooling: (1) ’max pooling’ with kernel size 2,2 and stride 2,2 partitions the 4x4
input feature map into 4 non-overlapping rectangles and returns the maximum for each rectangle. (2)
’average pooling’ with kernel size 2,2 and stride 1,1 partitions the 4x4 same input feature map into 9
in some cases overlapping rectangles and returns the average of each rectangle.

Nonlinear layer: Rectified Linear Unit (ReLU) layer

A CNN has to be able to approximate nonlinear functions. Thus, it needs at least one nonlinear layer. ReLU layers
have become a common approach since they can be computed very quickly.
Fully connected layer
In a CNN, the last layer is usually a fully connected layer. This layer is similar to the hidden layers of an MLP. It
takes the output of the previous layer, feature maps of high level features. Then, based on these features, a class is
chosen. For example, if the image shows a dog (class: dog), there might be feature maps that represent high level
features such as paws, snouts, or fur. These feature maps are created automatically.
Basic setup
A general deep learning network may be very deep and include complicated layers. An illustrative example for a
complete CNN is given in figure 3.15, where we show the simple network used for OCR classification in HALCON.
The hidden layers are
Convolutional –> ReLu, Pooling, –> Convolutional –> ReLu, Pooling –> Fully Connected
Each layer uses the output of the previous layer (see figure 3.15). The first hidden layers detect low level features
like edges and curves. The feature maps of the following layers describe higher level features. Filters deeper in the
network perceive information from a larger area of the original image. Finally, a fully connected layer is used to
compute the output. This way, the network is able to predict the class of an object.

Input

H
Output

Convolutional ReLu, pool. Convolutional ReLu, pool. Fully Connected

Figure 3.15: Schema of the convolutional neural network used in HALCON for OCR.
 Decisions to Make D-27

Chapter 4

Decisions to Make

This section gives you some hints how to select a suitable classification approach (section 4.1), the suitable features
that build the feature vectors (section 4.2), and the suitable training samples (section 4.3 on page 29). Note that
only some hints but no absolute rules can be given for almost all decisions that are related to classification, as the
best suited approach, features, and samples depend strongly on the specific application.

4.1 Select a Suitable Classification Approach

Select Approach
In most cases, we recommend to use either a DL, MLP, SVM, GMM, or k-NN classifier, as these classification
approaches are the most powerful and flexible ones. In table 4.1, the characteristics of these five classification
approaches are put together in a very brief way.
Based on the requirements and restrictions imposed by your application, you can use table 4.1 to select the best
suited classification approach. If you are not satisfied with the quality of the classification results, it is typically
not because of the chosen classifier but because of the used features or because of the quality and amount of the
training samples. Only if you are sure that the training data describes all the relevant characteristics of the objects
to be classified, it is worth to test if another classifier may produce better results.
For image segmentation, the four classification approaches MLP, SVM, GMM, and k-NN can be sped up signif-
icantly using a look-up table (see section 6.1.7 on page 70). But note that the so-called LUT-accelerated classifi-
cation is only suitable for images with a maximum of three channels. Furthermore, LUT-accelerated classification
leads to a slower offline phase and to higher memory requirements.

• You probably want to use the CNN deep learning classifier when you need a method able to achieve very
high accuracy and/or it is difficult to define the features necessary for your image classification problem. For
this approach, in comparison to our other classifiers, you do not need to define the features manually. Instead,
the training algorithm uses the images you already labeled (and therewith you assigned these images a class).
With this data the algorithm carries out transfer learning (see e.g., section 4.3 on page 29 or the reference
manual chapter “Deep Learning”), thus adapts an existing neural network for your specific application. More
data should help the training algorithm to train the network better. This means, the network should generalize
better from the given samples to a general case concerning your specific classification task. On the flip side
the training will take longer.
• The MLP classifier is especially well suited for applications that require a fast classification but allow for a
slow offline training phase. The complete training data should be available right from the beginning because
otherwise the time consuming training must be repeated from scratch. MLP classification does not support
novelty detection.
• The SVM classifier may often be tuned to achieve a slightly higher classification quality than the other
classifiers. But the classification speed is typically significantly slower than that of the MLP classifier. The
training of the SVM classifier is substantially faster than that of the MLP classifier, but it is typically too
slow for being used in the online phase. The SVM classifier requires significantly more memory than the
MLP classifier, while it requires less memory than the k-NN classifier. Typically, the memory requirements
rise with the number of training samples, i.e., for classification tasks with a huge number of training samples,
like OCR, the SVM classifier may become very large.
D-28 Decisions to Make

DL1 MLP SVM GMM k-NN

Training speed slow slow medium fast fast
Classification speed system and fast medium fast medium
network
depending:
medium to
fast
Automatic feature yes no no no no
extraction
Highest classification low number low number low number low number low number
speed is reached for 2 of classes of hidden of support of classes of training
and small nodes and vectors 3 samples
network classes
Memory require- network low medium low high5
ments 4 depending:
medium to
high
Use of additional yes no not recom- not recom- yes
training data 6 mended mended
Suited for high di- yes yes yes no yes
mensional feature
spaces
Suited for novelty no no yes yes yes
detection
1 Regarding only deep-learning-based models of type classification
2 Besides having a low dimensional feature space
3 The number of support vectors can be reduced with reduce_class_svm or reduce_ocr_class_svm
4 After removing the training samples from the classifier
5 The training samples cannot be removed from the k-NN classifier
6 Use of additional training data is possible without the need to retrain the whole classifier from scratch. Note, depending on the method you

may still use your whole dataset, see e.g., the reference manual entry for “Deep Learning”.

Table 4.1: Comparison of the characteristics of the four classifiers MLP, SVM, GMM, and k-NN.

• The GMM classifier is very fast both in training and classification, especially if the number of classes is
low. It is also very well suited for novelty detection. However it is restricted to applications that do not
require a high dimensional feature space.
• The k-NN classifier is especially well suited to test various configurations of features and training data
because the training of a k-NN classifier is very fast and it has no restrictions concerning the dimensionality
of the feature space. Furthermore, the classifier can be extended with additional training data very quickly.
Note that the k-NN classification is typically slower than the MLP classification and it requires substantially
more memory, which might be prohibitive in some applications.
• The classifier based on a 2D histogram is suitable for the pixel-based image segmentation of two-channel
images. It provides a very fast alternative if a 2D feature vector is sufficient for the classification task.
• The hyperbox and Euclidean classifiers are suitable for feature vectors of low dimension, e.g., when ap-
plying a color classification for image segmentation. Especially for classes that are built by rather compact
clusters, they are very fast. Compared to a LUT-accelerated classification using MLP, SVM, GMM, or k-NN,
the storage requirements are low and the feature space can easily be visualized.

For OCR, it is recommended to first try the pretrained font Universal (see Solution Guide I,section 18.8 on page
222), which is based on CNNs (see section 3.7 on page 23), before you try any other OCR classification approach.

4.2 Select Suitable Features

For all our classification approaches (except for the deep-learning-based ones), you need to select features that are
suitable for a classification. These features strongly depend on the specific application and the objects that have
 4.3 Select Suitable Training Samples D-29

to be classified. Thus, no fixed rules for their selection can be provided. For each application, you have to decide
individually , which features describe the object best. Generally, the following features can be used for the different
classification tasks:

• For a general classification all types of features, i.e., region features as well as color or texture, can be used
to build the feature vectors. The feature vectors have to be explicitly built by feature values that are derived
with a set of suitable operators.
• For image segmentation, the pixel values of a multi-channel color or texture image are used as features.
Here, you do not have to explicitly extract the feature vectors as they are derived automatically by the
corresponding image segmentation operators from the color or texture image.
• For OCR, a restricted set of region features is used to build the feature vectors. Here, you do not have to
explicitly calculate the features but select the feature types that are implicitly and internally calculated by the
corresponding OCR specific operators. The dimension of the resulting feature vector is equal or larger than
the number of selected feature types, as some feature types lead to several feature values (see section 7.7 on
page 89 for the list of available features).

If your objects are described best by texture, you can follow different approaches. You can, e.g., create a texture
image by applying the operator texture_laws with different parameters and combining the thus obtained indi-
vidual channels into a single image, e.g., using compose6 for a texture image containing six channels. Another
common approach is to use, e.g., the operator cooc_feature_image to calculate texture features like energy, cor-
relation, homogeneity, and contrast. We refer to Solution Guide I, chapter 15 on page 161 for further information

Select Samples
about texture.
If your objects are described best by region features, you can use any of the operators that are described in the
Reference Manual in section Regions/Features. For OCR, the set of available region features is restricted to
the set of features introduced in section 7.7 on page 89.
HDevelop provides convenience procedures (see calculate_features) to calculate multiple features
with given properties like rotational invariance, etc. in just a few calls. Additionally, HALCON of-
fers functionality to select suitable features automatically using the operators select_feature_set_mlp,
select_feature_set_svm, select_feature_set_gmm, and select_feature_set_knn. If
you are not sure which features to chose, you can use the HDevelop example programs
hdevelop/Classification/Feature-Selection/auto_select_region_features.hdev and
hdevelop/Applications/Object-Recognition-2D/classify_pills_auto_select_features.hdev
as a starting point, which make use of both, the procedures and the automatic feature selection.

4.3 Select Suitable Training Samples

In section 1 on page 7 we learned that classification is reasonable in all cases where objects have similarities, but
within undefined variations. To learn the similarities and variations, the classifier needs representative samples.
That is, the samples should not only show the significant features of the objects to classify but should also show a
large variety of allowed deviations. That is, if an object is described by a specific texture, small deviations from the
texture that are caused, e.g., by noise, should be covered by the samples. Or if an object is described by a region
having a specific size and orientation, the samples should contain several objects that deviate from both ’ideal’
values within a certain tolerance. Otherwise, only objects that exactly fit to the ’ideal’ object are found in the later
classification. In other words, the classifier has no sufficient generalization ability.
Generally, for the training of a classifier a large amount of samples with a realistic set of variations for the calculated
features should be provided for every available class. Otherwise, the result of the later classification may be
unsatisfying as the unknown objects show deviations from the trained data that were not considered during training.
Note that when applying transfer learning for deep-learning-based classification of general features (see e.g., sec-
tion 4.3 or the reference manual chapter “Deep Learning”), many already labeled images have to be provided for
each class. This means the tricks described below cannot be recommended in general and their appropriateness
depends strongly on the specific case and goal.
If, for any reason, no sufficient number of samples can be provided, some tricks may help:
D-30 Decisions to Make

• One trick is to generate artificial samples by copying the few available samples and slightly modifying them.
The modifications depend on the object to classify and the features used to find the class boundaries. When
working with texture images, e.g., noise can be added to slightly modify the copies of the samples. Or given
the example with the objects of a specific size and orientation, you can modify copies of the samples by, e.g.,
slightly changing their size using an erosion or dilation. And you can change their orientation by rotating
the image by different, but small angles. Ideally, you create several copies and modify them so that several
deviations in all allowed directions are covered.
• A second trick can be applied if the number of samples is unequally distributed for the different classes. For
example, you want to apply classification for quality inspection and you have a large amount of samples
for the good objects, but only a few samples for each of several error classes. Then, you can split the
classification task into two classification tasks. In the first instance, you merge all error classes into one
class, i.e., you have reduced the multi-class problem to a two-class problem. You have now a class with
good objects and the rejection class contains all erroneous objects, which in the sum are represented by a
larger number of samples. Then, if the type of error attached to the rejected objects is of interest, you apply
a second classification, this time without the lot of good examples. That is, you only use the samples of the
different error classes for the training and classify the objects that were rejected during the first classification
into one of these error classes.
 Classification of General Features D-31

Chapter 5

Classification of General Features

This section shows how to apply the different classifiers. The classification approaches implemented in HALCON
can be divided into two major groups. The first group consists of the methods MLP, SVM, GMM, and k-NN, where
the distinguishing features of each class have to be specified. The second group is given by deep-learning-based
methods (DL), where the network is trained by considering the inputs and outputs directly. Accordingly, the basic
approach for a classification with HALCON depends on the method group.
For the first group, thus MLP, SVM, GMM, and k-NN, it is possible to apply the classifier on arbitrary objects like
pixels or regions due to arbitrary features like color, texture, shape, or size. Here, pixels as well as regions can
be classified, in contrast to the image segmentation approach described in section 6 on page 57, which classifies
only pixels, or the OCR approach in section 7 on page 75, which classifies regions with focus on optical character
recognition. For all operators of this group, the general approach for a classification of arbitrary features, i.e., the
sequence of operators used, is similar. This approach is illustrated in section 5.1 by an example, which checks
the quality of halogen bulbs using shape features. In section 5.2, the steps of a classification and the involved
operators are listed for a brief overview. The parameters used for the operators are in many cases specific to the
individual approach because of the different underlying algorithms (see section 3 on page 15 for the theoretical

General Features
background). They are introduced in more detail in section 5.3 for MLP, section 5.4 for SVM, section 5.5 for
GMM, and section 5.6 for k-NN.
The second group consists of general deep-learning-based classification, where the classifier is applied to images.
The ’features’ are not hand-picked as for the other approaches, but chosen automatically during training. The
general approach is described in the chapter “Deep Learning . Model” and the workflow in the chapter “Deep
Learning . Classification”. A list of possible model parameters and an explanation to them is given in the reference
manual entry of get_dl_model_param.

5.1 General Approach (Classification of Arbitrary Features)

The general approach is similar for MLP, SVM, GMM, and k-NN classification (see figure 5.1). In all cases, a
classifier with specific properties is created. Then, known objects are investigated, i.e., you extract the features of
objects for which the classes are known and add the feature vectors together with the corresponding known class
ID to the classifier. With a training, the classifier then derives the rules for the classification, i.e., it decides how to
separate the classes from each other. To investigate unknown objects, i.e., to classify them, you extract the same
set of features for them that was used for the training, and classify the feature vectors with the trained classifier.
In the following, we illustrate the general approach with the example %HALCONEXAMPLES%\solution_guide\
classification\classify_halogen_bulbs.hdev. Here, halogen bulbs are classified into good, bad, and not
existent halogen bulbs (see figure 5.2). For that, the regions representing the insulation of the halogen bulbs are
investigated. The classification is applied with the SVM approach. The operator names for the MLP, GMM, and
k-NN classification differ only in their ending. That is, if you want to apply an MLP, GMM, or k-NN classification,
you mainly have to replace the ’svm’ by ’mlp’, ’gmm’, or ’knn’ in the corresponding operator names and adjust
different parameters. The parameters and their selection are described in section 6.1.3 for MLP, section 6.1.4 for
SVM, section 6.1.5 for GMM, and section 6.1.6 for k-NN.
The program starts with the assignment of the available classes. The halogen bulbs can be classified into the classes
’good’ (halogen bulb with sufficient insulation), ’bad’ (halogen bulb with insufficient insulation), or ’none’ (no
halogen bulb can be found in the image).
D-32 Classification of General Features

Create Classifier

Investigate Known Objects

Extract Features −> Feature Vectors

Assign Feature Vectors to Classes by Knowledge −> Samples
Add Samples to Classifier

Train Classifier

Investigate Unknown Objects

Extract Features −> Feature Vectors

Assign Feature Vectors to Classes by Classification

Figure 5.1: The basic steps of a general classification.

Figure 5.2: Classifying halogen bulbs into (from left to right): good, bad, and not existent halogen bulbs.

ClassNames := ['good', 'bad', 'none']

As the first step of the actual classification, an SVM classifier is created with the operator create_class_svm.
The returned handle of the classifier SVMHandle is needed for all classification specific operators that are applied
afterwards.

Nu := 0.05
KernelParam := 0.02
create_class_svm (7, 'rbf', KernelParam, Nu, |ClassNames|, 'one-versus-one', \
'principal_components', 5, SVMHandle)

As each classification application is unique, the classifier has to be trained for the current application. That is, the
rules for the classification have to be derived from a set of samples. In case of the SVM approach, e.g., the training
determines the optimal support vectors that separate the classes from each other (see section 3.4 on page 19).
A sample is an object for which the class membership is known. Generally, each kind of object can be classified
with the general classification approach as long as it can be described by a set of features or respectively the
feature’s values. Common objects for image processing are regions, pixels, or a combination of both. For the
example with the halogen bulbs, the objects that have to be trained and classified are the regions that represent
insulations of halogen bulbs. For each known object, the feature vector, which consists of values that are derived
from the extracted region, and the corresponding (known) class name build the training sample. For each class a
representative set of training samples must be provided to achieve suitable class boundaries. In the example, the
samples are added within the procedure add_samples_to_svm.
Within the procedure, for each class the corresponding images are obtained. Note that different methods can be
used to assign the class memberships for the objects of an image. In the example described in section 2 on page 11,
a tuple was used to assign the class name for each image. There, the sequence of the images and the sequence of
the elements in the tuple had to correspond. Here, the images of each class are stored in a directory that is named
 5.1 General Approach (Classification of Arbitrary Features) D-33

like the class. Thus, the procedure add_samples_to_svm uses the directory names to assign the feature vectors to
the classes. For example, the images containing the good halogen bulbs are stored in the directory ’good’. Then,
for each class, all images are read from the corresponding directory.
Now, for each image the region of the halogen bulb’s insulation is extracted by the operator threshold, the
features of the region are extracted inside the procedure calculate_features, and the feature vector is added
together with the corresponding class ID to the classifier using the operator add_sample_class_svm.

procedure add_samples_to_svm (ClassNames, SVMHandle, WindowHandle, ReadPath):::

for ClassNumber := 0 to |ClassNames| - 1 by 1
list_files (ReadPath + ClassNames[ClassNumber], 'files', Files)
Selection := regexp_select(Files,'.*[.]png')
for Index := 0 to |Selection| - 1 by 1
read_image (Image, Selection[Index])
threshold (Image, Region, 0, 40)
calculate_features (Region, Features)
add_sample_class_svm (SVMHandle, Features, ClassNumber)
endfor
endfor
return ()

The feature vectors that are used to train the classifier and those that are classified for new objects must con-
sist of the same set of features. In the example program, the features are calculated inside the procedure
calculate_features and comprise

• the ’area’ of the region,

• the ’compactness’ of the region,

• the four geometric moments (’PSI1’, ’PSI2’, ’PSI3’, and ’PSI4’) of the region, which are invariant to
translation and general linear transformations, and

General Features
• the ’convexity’ of the region.

Note that feature vectors have to consist of real values. As some of the calculated features are described by integer
values, e.g., the feature ’area’, which corresponds to the number of pixels contained in a region, the feature
vector is transformed into a tuple of real values before it is added to the classifier.

procedure calculate_features (Region, Features)

area_center (Region, Area, Row, Column)
compactness (Region, Compactness)
moments_region_central_invar (Region, PSI1, PSI2, PSI3, PSI4)
convexity (Region, Convexity)
Features := real([Area,Compactness,PSI1,PSI2,PSI3,PSI4,Convexity])
return ()

After adding all samples to the classifier with the procedure add_samples_to_svm, the actual training is applied
with the operator train_class_svm. In this step, the classifier derives its classification rules.

train_class_svm (SVMHandle, 0.001, 'default')

These classification rules are applied now inside the procedure classify_regions_with_svm to halogen bulbs
of unknown classes. The procedure works similar as the procedure for adding the training samples. But now, the
images that contain the unknown types of halogen bulbs are read, no class information is available, and instead
of adding samples to the classifier, the operator classify_class_svm is applied to classify the unknown feature
vectors with the derived classification rules.
D-34 Classification of General Features

procedure classify_regions_with_svm (SVMHandle, Colors, ClassNames,

ReadPath):::
list_files (ReadPath, ['files', 'recursive'], Files)
Selection := regexp_select(Files,'.*[.]png')
read_image (Image, Selection[0])
for Index := 0 to |Selection| - 1 by 1
read_image (Image, Selection[Index])
threshold (Image, Region, 0, 40)
calculate_features (Region, Features)
classify_class_svm (SVMHandle, Features, 1, Class)
endfor
return ()

The example shows the application of operators that are essential for a classification. Further operators are provided
that can be used, e.g., to separate the training from the classification. That is, you run a program that applies the
training offline, save the trained classifier to file with write_class_svm, and in another program you read the
classifier from file again with read_class_svm to classify your data in an online process. When closing the
training program, the samples are not stored automatically. To store them to file for later access you apply the
operator write_samples_class_svm. A later access using read_samples_class_svm may be necessary, e.g.,
if you want to repeat the training with additional training samples.
The following sections provide you with a list of involved operators (section 5.2) and go deeper into the specific pa-
rameter adjustment needed for MLP (section 5.3), SVM (section 5.4), GMM (section 5.5), and k-NN (section 5.6).

5.2 Involved Operators (Overview)

This section gives a brief overview on the operators that are provided for a general MLP, SVM, GMM, and k-NN
classification. First, the operators for the basic steps of a classification are introduced in section 5.2.1. Then, some
advanced operators are introduced in section 5.2.2. The following sections introduce the individual parameters for
the basic operators and provide tips for their adjustment.
DL classification is implemented within the more general deep learning model. For the general workflow we refer
to “Deep Learning . Classification”, mentioning the important steps with their involved operators and helpful
procedures.

5.2.1 Basic Steps: MLP, SVM, GMM, and k-NN

Summarizing the information obtained in section 5.1 on page 31, the classification consists of the following basic
steps and operators, which are applied in the same order as listed here:

1. Create a classifier. Here, some important properties of the classifier are defined. The returned handle is
needed in all later classification steps.
• create_class_mlp
• create_class_svm
• create_class_gmm
• create_class_knn
2. Predefine the sequence in which the classes are defined and later accessed, i.e., define the correspondences
between the class IDs and the class names. This step may as well be applied before the creation of the
classifier.
3. Get feature vectors for sample objects of known class IDs. The operators that are suitable to obtain the
features depend strongly on the specific application and thus are not part of this overview.
4. Successively add samples, i.e., feature vectors and their corresponding class IDs to the classifier.
• add_sample_class_mlp
• add_sample_class_svm
 5.2 Involved Operators (Overview) D-35

• add_sample_class_gmm
• add_sample_class_knn
5. Train the classifier. Here, the boundaries between the classes are derived from the training samples.
• train_class_mlp
• train_class_svm
• train_class_gmm
• train_class_knn
6. Store the used samples to file and access them in a later step (optionally).
• write_samples_class_mlp and read_samples_class_mlp
• write_samples_class_svm and read_samples_class_svm
• write_samples_class_gmm and read_samples_class_gmm
• Note that there are no operators for writing and reading the samples of a k-NN classifier separately
because the samples are an intrinsic component of the k-NN classifier. Use write_class_knn and
read_class_knn instead.
7. Store the trained classifier to file and read it from file again.
• write_class_mlp (default file extension: .gmc) and read_class_mlp
• write_class_svm (default file extension: .gsc) and read_class_svm
• write_class_gmm (default file extension: .ggc) and read_class_gmm
• write_class_knn (default file extension: .gnc) and read_class_knn
Note that the samples cannot be deleted from a k-NN classifier because they are an intrinsic component
of this classifier.
8. Get feature vectors for objects of unknown class. These feature vectors have to contain the same features (in

General Features
the same order) that were used to define the training samples.
9. Classify the new feature vectors. That is, insert the new feature vector to one of the following operators and
get the corresponding class ID.
• classify_class_mlp
• classify_class_svm
• classify_class_gmm
• classify_class_knn

5.2.2 Advanced Steps: MLP, SVM, GMM, and k-NN

This section mentions advanced operators, which can be applied for possible additional steps. Especially if the
training and classification do not lead to a satisfying result, it is helpful to access some information that is implicitly
contained in the model. Available steps to query information are:

• Access an individual sample from the training data. This is needed, e.g., to check the correctness of
its class assignment. The sample had to be stored previously by the operator add_sample_class_mlp,
add_sample_class_svm, add_sample_class_gmm, or add_sample_class_knn, respectively.
– get_sample_class_mlp
– get_sample_class_svm
– get_sample_class_gmm
– get_sample_class_knn
• Get the number of samples that are stored in the training data. The obtained number is needed, e.g., to access
the individual samples or to know how much individual samples you can access.
– get_sample_num_class_mlp
D-36 Classification of General Features

– get_sample_num_class_svm
– get_sample_num_class_gmm
– get_sample_num_class_knn

• Get information about the content of the preprocessed feature vectors. This information is reasonable, if the
parameter Preprocessing was set to ’principal_components’ or ’canonical_variates’ during the
creation of the classifier. Then, you can check if the information that is contained in the preprocessed feature
vector still contains significant data or if a different preprocessing parameter, e.g., ’normalization’, is to
be preferred.

– get_prep_info_class_mlp
– get_prep_info_class_svm
– get_prep_info_class_gmm
– Note that this information cannot be retrieved from a k-NN classifier because the respective kind of
preprocessing is not available for k-NN classifiers.
• Get the parameter values that were set during the creation of the classifier. This is needed if the offline
training and the online classification are separated and the information about the training part is not available
anymore.
– get_params_class_mlp
– get_params_class_svm
– get_params_class_gmm
– get_params_class_knn

Furthermore, there are operators that are available only for specific classifiers:

• For MLP and GMM you can evaluate the probabilities of a feature vector to belong to a specific class. That
is, you can determine the probabilities for each available class and not only for the most probable classes. If
only the most probable classes are of interest, no explicit evaluation is necessary, as these probabilities are
returned also when classifying the feature vector.
– evaluate_class_mlp
– evaluate_class_gmm
• For SVM you can reduce the number of support vectors returned by the offline training to speed up the
following online classification.

– reduce_class_svm
• Additionally, for SVM the number or index of the support vectors can be determined after the training. This
is suitable for the visualization of the support vectors and for diagnostic reasons.
– get_support_vector_num_class
– get_support_vector_class
• For k-NN you can set various parameters that control the behavior of the classifier with the operator
– set_params_class_knn.

This includes the number k of nearest neighbors, the kind of result that is to be returned by the classifier, and
parameters that control the trade-off between quality and speed of the classification.
 5.3 Parameter Setting for MLP D-37

5.3 Parameter Setting for MLP

This section goes deeper into the parameter adjustment for an MLP classification. We recommend to first adjust
the parameters so that the classification result is satisfying. The most important parameter that has to be adjusted
to get the MLP classifier work optimally is

• NumHidden (create_class_mlp).

If the classification generally works, you can start to tune the speed. The most important parameters to enhance
the speed are

• Preprocessing / NumComponents (create_class_mlp) and

• MaxIterations (train_class_mlp).

In the following, we introduce the parameters of the basic MLP operators. The focus is on the parameters for
which the setting is not immediately obvious or for which it is not immediately obvious how they influence the
classification. These are mainly the parameters needed for the creation and training of the classifier. Further
information about these operators as well as the usage of the operators with obvious parameter settings can be
found in the Reference Manual entries for the individual operators.

5.3.1 Adjusting create_class_mlp

An MLP classifier is created with the operator create_class_mlp. There, several properties of the classifier are
defined that are important for the following classification steps. The returned handle is needed (and modified) in
all following steps. The following parameters can be adjusted:

Parameter NumInput

General Features
The input parameter NumInput specifies the dimension of the feature vectors used for the training as well as for
the classification. Opposite to the GMM classifier (see section 5.5 on page 47), a number of 500 features is still
realistic.

Parameter NumHidden
The input parameter NumHidden defines the number of units of the hidden layer of the multi-layer neural net (see
section 3.3 on page 18) and significantly influences the result of the classification and thus should be adjusted very
carefully. Its value should be in a similar value range as NumInput and NumOutput. Smaller values lead to a less
complex separating hyperplane, but in many cases nevertheless may lead to good results. With a very large value
for NumHidden, you run the risk of overfitting (see figure 5.3). That is, the classifier uses unimportant details like
noise to build the class boundaries. That way, the classifier works very well for the training data, but fails for
unknown feature vectors that do not contain the same unimportant details. In other words, overfitting means that
the classifier looses its generalization ability.
To adjust NumHidden, it is recommended to apply tests with independent test data, e.g., using the cross
validation introduced in section 8.1 on page 93. Note that the example %HALCONEXAMPLES%\hdevelop\
Classification\Neural-Nets\class_overlap.hdev provides further hints about the influence of different
values for NumHidden.

Parameter NumOutput
The input parameter NumOutput specifies the number of classes.

Parameter OutputFunction
The input parameter OutputFunction describes the functions used by the output unit of the neural net. Available
values are ’softmax’, ’logistic’, and ’linear’. In almost all classification applications, OutputFunction
should be set to ’softmax’. The value ’logistic’ can be used for classification problems with multiple in-
dependent logical attributes as output, but this kind of classification problems is very rare in practice. The value
’linear’ is used for least squares fitting (regression) and not for classification. Thus, you can ignore it here.
D-38 Classification of General Features

Error

NumHidden

Underfitting Ideal Value Overfitting

Range

Figure 5.3: The value of NumHidden should be adjusted carefully to avoid under- or overfitting (note that the illus-
trated curve is idealized and in practice would be less straight).

Parameters Preprocessing /NumComponents

The input parameter Preprocessing defines the type of preprocessing applied to the feature vector for the training
as well as later for the classification or evaluation. A preprocessing of the feature vector can be used to speed up
the training as well as the classification. Sometimes, even the recognition rate can be enhanced.
Available values are ’none’, ’normalization’, ’principal_components’, and ’canonical_variates’. In
most cases, the preprocessing should be set to ’normalization’ as it enhances the speed without loosing relevant
information compared to using no preprocessing (’none’). The feature vectors are normalized by subtracting the
mean of the training vectors and dividing the result by the standard deviation of the individual components of the
training vectors. Hence, the transformed feature vectors have a mean of 0 and a standard deviation of 1. The
normalization does not change the length of the feature vector.
If speed is important and your data is expected to be highly correlated, you can reduce the dimension of the
feature vector using a principal component analysis (’principal_components’). There, the feature vectors are
normalized and additionally transformed such that the covariance matrix becomes a diagonal matrix. Thus, the
amount of data can be reduced without losing a large amount of information.
If you know that your classes are linearly separable, you can also use canonical variates
(’canonical_variates’). This approach is known also as linear discriminant analysis. There, the trans-
formation of the normalized feature vectors decorrelates the training vectors on average over all classes. At the
same time, the transformation maximally separates the mean values of the individual classes. This approach
combines the advantages of a principal component analysis with an optimized separability of the classes after the
data reduction. But note that the parameter ’canonical_variates’ is recommended only for linearly separable
classes. For MLP, ’canonical_variates’ can only be used if OutputFunction is set to ’softmax’.
Figure 5.4 and figure 5.5 illustrate how ’principal_components’ and ’canonical_variates’, dependent
on the distribution of the feature vectors, can reduce the feature vectors to a lower number of components by
transforming the feature space and projecting the feature vectors to one of the principal axes.
The input parameter NumComponents defines the number of components to which the feature vector is reduced if
a preprocessing is selected that reduces the dimension of the feature vector. In particular, NumComponents has to
be adjusted only if Preprocessing is set to ’principal_components’ or ’canonical_variates’.
If Preprocessing is set to ’principal_components’ or ’canonical_variates’ you can use the operator
get_prep_info_class_mlp to check if the content of the transformed feature vectors still contain significant
data. Furthermore, you can use the operator to determine the optimum number of components. Then, you first
create a test classifier with, e.g., NumComponents set to NumInput, generate and add the training samples to the
classifier, and then apply get_prep_info_class_mlp. The output parameter CumInformationCont is a tuple
containing numbers between 0 and 1. These numbers describe the amount of the original data that is covered by
the transformed data. That is, if you want to have at least 90% of the original data covered by the transformed
data, you search for the first value that is larger than 0.9 and use the corresponding index number of the tuple
CumInformationCont as value for a new NumComponents. Then, you create a new classifier for the final training
 5.3 Parameter Setting for MLP D-39

Feature 2: Feature 2:
principal components

x x
x
xx

x xx x

x x xx
x x

x
xx

x x
x
xx
0 0

x x
x
x
x

x
x

Feature 1: Feature 1:
0 0

Figure 5.4: After transforming the feature space via principal component analysis, the illustrated linearly separable
classes can be separated using only one feature.

Feature 2: Feature 2: Feature 2:

principal components canonical variates

x x x
x x x xx

x x
x x
x

x x x

xx
x xx
x

x
x

x
x x

x
x
x x

0 0 0
x

x x x
x x
x

xx x
x

x
x
x

x
x

Feature 1: Feature 1: Feature 1:

0 0 0

Figure 5.5: Here, after transforming the feature space via principal component analysis, the illustrated linearly
separable classes still need two features to be separated. After transforming the feature space using
canonical variates they can be separated by a single feature.

General Features
(this time NumComponents is set to the new value). Note that it is suitable to store the samples into a file during
the test training (write_samples_class_mlp) so that you do not have to successively add the training samples
again but simply read in the sample file using read_samples_class_mlp.

Parameter RandSeed
The weights of the MLP (see section 3.3 on page 18) are initialized by a random number. For the sake of repro-
ducibility the seed value for this random number is stored in the input parameter RandSeed.

Parameter MLPHandle
The output parameter of create_class_mlp is the MLPHandle, which is needed for all following classification
specific operators.

5.3.2 Adjusting add_sample_class_mlp

A single sample is added to the MLP classifier using add_sample_class_mlp. For the training, several sam-
ples have to be added by successively calling add_sample_class_mlp with different samples. The following
parameters can be adjusted:

Parameter MLPHandle
The input and output parameter MLPHandle is the handle of the classifier that was created with
create_class_mlp and to which the samples subsequently were added with add_sample_class_mlp. After
applying add_sample_class_mlp for all available samples, the handle is prepared for the actual training of the
classifier.
D-40 Classification of General Features

Parameter Features
The input parameter Features contains the feature vector of a sample to be added to the classifier with
add_sample_class_mlp. This feature vector is a tuple of values. Each value describes a specific numeric feature.
Note that the feature vector must consist of real numbers. If you have integer numbers, you have to transform them
into real numbers. Otherwise, an error message is raised.

Parameter Target
The input parameter Target describes the target vector, i.e., you assign the corresponding class ID to the feature
vector.
If OutputFunction is set to ’softmax’, the target vector is a tuple that contains exactly one element with the
value 1 and several elements with the value 0. The size of the vector corresponds to the number of available classes
specified by NumOutput inside create_class_mlp. The index of the element with the value 1 defines the class
the feature vector Features belongs to. Alternatively, a single integer containing the class number (counted from
0) can be specified.
For OutputFunction set to ’logistic’, the target vector consists of values that are either 0 or 1. Each 1 shows
that the corresponding feature is present.
If OutputFunction is set to ’linear’, the target vector can contain arbitrary real numbers. As this parameter
value is used for least squares fitting (regression) and not for classification, it is not explained further.

5.3.3 Adjusting train_class_mlp

The training of the MLP classifier is applied with train_class_mlp. Training the MLP means to determine the
optimum values of the MLP weights (see section 3.3 on page 18). For this, a sufficient number of training samples
is necessary. Training is performed by a complex nonlinear optimization process that minimizes the discrepancy of
the MLP output and the target vectors that were defined with add_sample_class_mlp. The following parameters
can be adjusted:

Parameter MLPHandle
The input and output parameter MLPHandle is the handle of the classifier that was created with
create_class_mlp and for which samples were stored either by adding them via add_sample_class_mlp or by
reading them in with read_samples_class_mlp. After applying train_class_mlp, the handle is prepared for
the actual classification of unknown data. That is, it then contains information about how to separate the classes.

Parameters MaxIterations / WeightTolerance / ErrorTolerance

The input parameters MaxIterations, WeightTolerance, and ErrorTolerance control the nonlinear opti-
mization algorithm. MaxIterations specifies the number of iterations of the optimization algorithm. The op-
timization is terminated if the weight change is smaller than WeightTolerance and the change of the error is
smaller than ErrorTolerance. In any case, the optimization is terminated after at most MaxIterations itera-
tions. For the latter, values between 100 and 200 are sufficient in most cases. The default value is 200. By reducing
this value, the speed of the training can be enhanced. For the parameters WeightTolerance and ErrorTolerance
the default values do not have to be changed in most cases.

Parameter Error
The output parameter Error returns the error of the MLP with the optimal weights on the training samples.

Parameter ErrorLog
The output parameter ErrorLog returns the error value as a function of the number of iterations. This function can
be used to decide if a second training with the same training samples but a different value for RandSeed should be
applied, which is the case if the function runs into a local minimum.
 5.3 Parameter Setting for MLP D-41

5.3.4 Adjusting evaluate_class_mlp

The operator evaluate_class_mlp can be used to evaluate the probabilities for a feature vector to belong to each
of the available classes. If only the probabilities for the two classes to which the feature vector most likely belongs
are searched for, no evaluation is necessary, as these probabilities are returned also for the final classification of the
feature vector. The following parameters can be adjusted:

Parameter MLPHandle
The input parameter MLPHandle is the handle of the classifier that was previously trained with the operator
train_class_mlp.

Parameter Features
The input parameter Features contains the feature vector that is evaluated. The feature vector must consist of the
same features as the feature vectors used for the training samples within add_sample_class_mlp.

Parameter Result
The output parameter Result returns the result of the evaluation. This result has different meanings, dependent on
the OutputFunction that was set with create_class_mlp. If OutputFunction was set to ’softmax’, which
should be the case for most classification applications, the returned tuple consists of probability values. Each value
describes the probability of the given feature vector to belong to the corresponding class. If OutputFunction
was set to ’logistic’, the elements of the returned tuple represent the presences of the respective independent
attributes.

5.3.5 Adjusting classify_class_mlp

The operator classify_class_mlp classifies a feature vector according to the class boundaries that were derived

General Features
during the training. It can only be called if OutputFunction was set to ’softmax’ in create_class_mlp. The
following parameters can be adjusted:

Parameter MLPHandle
The input parameter MLPHandle describes the handle that was created with create_class_mlp, to which samples
were added with add_sample_class_mlp, and that was trained with train_class_mlp. The handle contains
all the information that the classifier needs to assign an unknown feature vector to one of the available classes.

Parameter Features
The input parameter Features contains the feature vector of the object that is to be classified. The fea-
ture vector must consist of the same features as the feature vectors used for the training samples within
add_sample_class_mlp.

Parameter Num
The input parameter Num specifies the number of best classes to be searched for. Generally, Num is set to 1 if only
the class with the best probability is searched for, and to 2 if the second best class is also of interest, e.g., because
the classes overlap.

Parameter Class
The output parameter Class returns the result of classifying the feature vector with the trained MLP classifier, i.e.,
a tuple containing Num elements. That is, if Num is set to 1, a single value is returned that corresponds to the class
with the highest probability. If Num is set to 2, the first element contains the class with the highest probability and
the second element contains the second best class.
D-42 Classification of General Features

Parameter Confidence
The output parameter Confidence outputs the confidence of the classification. Note that in comparison to the
probabilities returned for a GMM classification, here the returned values can be influenced by outliers, which is
caused by the specific way an MLP is calculated. For example, the confidence may be high for a feature vector that
is far from the rest of the training samples of the specific class but significantly on the same side of the separating
hypersurface. On the other side, the confidence can be low for objects that are significantly within the cluster of
training samples of a specific class but near to the separating hypersurface as two classes overlap at this part of the
cluster (see figure 5.6).
High Confidence

x x
x x xx x
x
x x xx
x x
x xx x x

x
x

x
x
x xx
x x x
x x

Hy
pe
x

rp
lan
Low Confidence e

Figure 5.6: The confidence calculated for an MLP classification can be unexpected for outliers. That is, feature
vectors that are far away from their class centers can be classified with high confidence or those that
are near to the class centers but inside an overlapping area between two classes can be classified with
low confidence.

5.4 Parameter Setting for SVM

This section goes deeper into the parameter adjustment for an SVM classification. We recommend to first adjust
the parameters so that the classification result is satisfying. The most important parameters that have to be adjusted
to get the SVM classifier work optimally are:

• Nu (create_class_svm)
• KernelParam (create_class_svm)

If the classification generally works, you can start to tune the speed. The most important parameters to enhance
the speed are:

• Preprocessing (create_class_svm):
A combination of KernelType set to ’rbf’ and the Preprocessing set to ’principal_components’
speeds up the SVM significantly (even faster than MLP), as the features are reduced and thus the dimension
for the support vectors are reduced as well.
• MaxError (reduce_class_svm)

In the following, we introduce the parameters of the basic SVM operators. The focus is on the parameters for
which the setting is not immediately obvious or for which it is not immediately obvious how they influence the
classification. These are mainly the parameters needed for the creation and training of the classifier. Further
information about these operators as well as the usage of the operators with obvious parameter settings are provided
in the Reference Manual entries for the individual operators.

5.4.1 Adjusting create_class_svm

An SVM classifier is created with the operator create_class_svm. There, several properties of the classifier are
defined that are important for the following classification steps. The returned handle is needed (and modified) in
all following steps. For the operator, the following parameters can be adjusted:
 5.4 Parameter Setting for SVM D-43

Parameter NumFeatures
The input parameter NumFeatures specifies the dimension of the feature vectors used for the training as well as
for the classification. Opposite to the GMM classifier (see section 5.5 on page 47), a number of 500 features is still
realistic.

Parameters KernelType / KernelParam

In section 3.4 on page 19 we saw for an SVM classification that the feature space is transformed into a higher
feature space by a kernel to get linearly separable classes. The input parameter KernelType defines how the
feature space is mapped into this higher dimension. The mapping that is suitable and recommended in most cases
uses a kernel that is based on the Gauss error distribution curve and is called Gaussian radial basis function kernel
(’rbf’).
If KernelType is set to ’rbf’, the input parameter KernelParam is used to adjust the γ of the error curve (see
figure 5.7) and should be adjusted very carefully. If the value for γ is very high, the number of support vectors
increases, which results on one hand in an overfitting, i.e., the generalization ability of the classifier is lost (for
overfitting see also the description of NumHidden in section 5.3.1 on page 37) and on the other hand the speed is
reduced. On the other side, with a very low value for γ, an underfitting occurs, i.e., the number of support vectors
is not sufficient to obtain a satisfying classification result.
It is recommended to start with a small γ and then progressively increase it. Generally, it is recommended to
simultaneously search for a suitable Nu-γ pair, as these together define how complex the separating hypersurface
becomes. The search can be applied, e.g., using the cross validation that is described in section 8.1 on page 93.

General Features
Figure 5.7: γ describes the amount of influence of a support vector upon its surrounding.

Besides ’rbf’, you can also select a linear or polynomial kernel for KernelType, but these kernels should be
used only in very special cases (see below).
The linear kernel (KernelType set to ’linear’) transforms the feature space using a dot product. The linear
kernel should be used only if the classes are expected to be linearly separable. If a linear kernel is selected, the
parameter KernelParam has no meaning and can be ignored.
The polynomial kernels (KernelType set to ’polynomial_homogeneous’ or ’polynomial_inhomogeneous’)
in very rare cases can be used if the classification with ’rbf’ was not successful, but in most cases, ’rbf’ leads
to a better result. If a polynomial kernel is selected, the parameter KernelParam describes the degree ’d’ of the
polynomial. Note that a degree higher than 10 might result in numerical problems.

Parameter Nu
For classes that are not linearly separable, data from different classes may overlap. The input parameter Nu regu-
larizes the separation of the classes, i.e., with Nu, the upper bound for training errors within the overlapping areas
between the classes is adjusted (see figure 5.8) and at the same time the lower bound for the number of support
vectors is determined. Nu should be adjusted very carefully. Its value must be a real number between 0 and 1. As
a rule of thumb, it should be set to the expected error ratio of the specific dataset, e.g., to 0.05 when expecting a
maximum training error of 5%. The training error occurs because of, e.g., overlapping classes. Note that a very
large Nu results in a large dataset and thus reduces the speed significantly. Additionally, with a very large Nu the
training may be aborted and an error handling message is raised. Then, Nu has to be chosen smaller. On the other
hand, a very small Nu leads to instable numerics, i.e., many feature vectors would be classified incorrectly.
D-44 Classification of General Features

incorrectly classified
training data

Figure 5.8: The parameter Nu determines the amount of the incorrectly classified training data within the overlap
between two classes.

To select a suitable value for Nu, it is recommended to start with a small value and then progressively increase
it. Generally, it is recommended to simultaneously search for a suitable Nu-γ pair, as these together define how
complex the separating hypersurface becomes. The search can be applied, i.e., using the cross validation that is
described in section 8.1 on page 93.

Parameter NumClasses
The input parameter NumClasses specifies the number of classes.

Parameter Mode
As we saw in section 3.4 on page 19, SVM can handle only two-class problems. With the input parameter Mode
you define if your application is a two-class problem or if you want to extent the SVM to a multi-class problem.
Having a two-class problem, you have training data for a single class and decide during the classification if a
feature vector belongs to the trained class or not. That is, the hyperplane lies around the training data and implicitly
separates the training data from a rejection class. This Mode is called ’novelty-detection’ and can only be
applied if KernelType is set to ’rbf’.
If you want to extent the SVM to a multi-class problem, you have to divide the decision into binary sub-
cases. There, you have two possibilities. Either, you set Mode to ’one-versus-one’ or you set it to
’one-versus-all’.
When using the mode ’one-versus-one’, for each pair of classes a binary classifier is created and the class
that wins most comparisons is selected. Here, n classes result in n(n-1)/2 classifiers. This approach is fast but
suitable only for a small number of classes (approximately up to 10).
For ’one-versus-all’, each class is compared to the rest of the training data and the class with the maximum
distance to the hypersurface is selected. Here, the number of needed binary classifiers corresponds to the number
of classes. This approach is not as fast as ’one-versus-one’, but can and should be used for a higher number of
classes.

Parameters Preprocessing / NumComponents

The input parameter Preprocessing defines the type of preprocessing applied to the feature vector for the training
as well as later for the classification or evaluation. A preprocessing of the feature vector can be used to speed up
the training as well as the classification. Sometimes, even the recognition rate can be enhanced.
Available values are ’none’, ’normalization’, ’principal_components’, and ’canonical_variates’.
In most cases, the preprocessing should be set to ’normalization’ as it enhances the speed without loosing
relevant information compared to using no preprocessing (’none’). The values ’principal_components’ and
in rare cases ’canonical_variates’ can be used to enhance the speed. As the preprocessing types are the same
as used for an MLP classification, we refer to section 5.3.1 on page 38 for further information.
 5.4 Parameter Setting for SVM D-45

The input parameter NumComponents defines the number of components to which the feature vector is reduced
if a preprocessing is selected that reduces the dimension of the feature vector. In particular, NumComponents
has to be adjusted only if Preprocessing is set to ’principal_components’ or ’canonical_variates’.
If Preprocessing is set to ’principal_components’ or ’canonical_variates’ you can use the operator
get_prep_info_class_svm to determine the optimum number of components as described in section 5.3.1 on
page 38.

Parameter SVMHandle
The output parameter of create_class_svm is the SVMHandle, which is needed for all following classification
specific operators.

5.4.2 Adjusting add_sample_class_svm

A single sample is added to the SVM classifier using add_sample_class_svm. For the training, several sam-
ples have to be added by successively calling add_sample_class_svm with different samples. The following
parameters can be adjusted:

Parameter SVMHandle
The input and output parameter SVMHandle is the handle of the classifier that was created with
create_class_svm and to which the samples subsequently are added with add_sample_class_svm. After
applying add_sample_class_svm for all available samples, the handle is prepared for the actual training of the
classifier.

Parameter Features
The input parameter Features contains the feature vector of a sample to be added to the classifier with
add_sample_class_svm. This feature vector is a tuple of values. Each value describes a specific numeric feature.

General Features
Note that the feature vector must consist of real numbers. If you have integer numbers, you have to transform them
into real numbers. Otherwise, an error message is raised.

Parameter Class
The input parameter Class contains the ID of the class the feature vector belongs to. The ID is an integer number
between 0 and ’Number of Classes -1’. If you created a tuple with class names, the class ID is the index of
the corresponding class name in the tuple.

5.4.3 Adjusting train_class_svm

The training of the SVM classifier is applied with train_class_svm. The following parameters can be adjusted:

Parameter SVMHandle
The input and output parameter SVMHandle is the handle of the classifier that was created with
create_class_svm and for which samples were stored either by adding them via add_sample_class_svm or by
reading them in with read_samples_class_svm. After applying train_class_svm, the handle is prepared for
the actual classification of unknown data. That is, it then contains information about how to separate the classes.

Parameter Epsilon
Training the SVM means to gradually optimize the function that determines the class boundaries. This optimization
stops if the gradient of the function falls below a certain threshold. This threshold is set with the input parameter
Epsilon. In most cases it should be set to the default value, which is 0.001. With a too small threshold, the opti-
mization becomes slower without leading to a better recognition rate. With a too large threshold, the optimization
stops before the optimum is found, i.e., the recognition rate may be not satisfying.
There are two cases, in which changing the value of Epsilon might be reasonable. First, when having a very small
Nu and a small or unbalanced set of training data, it may be suitable to set Epsilon smaller than the default value
D-46 Classification of General Features

to enhance the resulting recognition rate. Second, when applying a cross validation to search for a suitable Nu-γ
pair (see section 8.1 on page 93), it is recommended to select a larger value for Epsilon during the search. Thus,
the cross validation is sped up without significantly changing the parameters for the optimal kernel. Having found
the optimal Nu-γ pair, the final training is applied again with the small (default) value.

Parameter TrainMode
The input parameter TrainMode determines the mode for the training. We recommend to use the mode ’default’
in most cases. There, the whole set of available samples is trained in one step. Appending a new set of samples to a
previously applied training is possible with the mode ’add_sv_to_train_set’. This mode has some advantages
that are listed in the Reference Manual entry for train_class_svm, but you have to be aware that only the support
vectors that resulted from the previously applied training are reused. The samples of the previously applied training
are ignored. This most likely leads to a different hypersurface than obtained with a training that uses all available
training samples in one step. The risk of obtaining a hypersurface that is not suitable for all available samples is
illustrated in figure 5.9.
New Training Samples

o
o
o

o
o

o o o

o
o

o o o
o

o o

o o
o o
o o

o
o
o

o o

o o
o o
o o

o
o

o
o

o
o

x x x
o

x x x
x

x
x
x

x
x

x x

x x
x x
x x

x x

x x
x x
x x

x x
x x

x
x

x
x
x

a) b) c) d)

Figure 5.9: Risk of appending a second training: a) training samples of the first training and the obtained hypersur-
face, b) new samples added for a second training, c) hypersurface obtained by a second training using
the new samples with the support vectors obtained by the first training, d) hypersurface obtained by a
new training that uses all available samples.

5.4.4 Adjusting reduce_class_svm

The operator reduce_class_svm can be used to reduce the number of support vectors that were returned by the
training. This is suitable to speed up the online classification. The following parameters can be adjusted:

Parameter SVMHandle
The input parameter SVMHandle is the handle of the classifier that was created with create_class_svm, to which
training samples were added with add_sample_class_svm, and which was trained with train_class_svm.
reduce_class_svm does not modify the handle but creates a copy of it (SVMHandleReduced) and modifies the
copy.

Parameters Method / MinRemainingSV / MaxError

The input parameter Method defines the method used to reduce the number of support vectors. Momentarily, only
the method ’bottom_up’ is available. There, the number of support vectors is reduced by iteratively merging the
support vectors until either the minimum number of support vectors that is set with MinRemainingSV is reached,
or until the accumulated maximum error exceeds the threshold that is set with MaxError.
Note that the approximation of the original support vectors by a reduced number of support vectors reduces also
the complexity of the hypersurface and thus can lead to a poor classification rate. A common approach is to start
with a small value for MaxError, e.g, 0.0001, and to increase it step by step. To control the reduction ratio, the
number of remaining support vectors is checked by get_support_vector_num_class and the classification rate
is checked by classifying a separate test data with classify_class_svm.

Parameter SVMHandleReduced
The output parameter SVMHandleReduced returns the copied and modified handle of a classifier that has the same
parametrization as the original handle but a different support vector expansion. Additionally, it does not contain
the training samples that are stored with the original handle.
 5.5 Parameter Setting for GMM D-47

5.4.5 Adjusting classify_class_svm

The operator classify_class_svm is used to decide to which of the trained classes an unknown feature vector
belongs. The following parameters can be adjusted:

Parameter SVMHandle
The input parameter SVMHandle describes the handle that was created with create_class_svm, to which samples
were added with add_sample_class_svm, and that was trained with train_class_svm. The handle contains
all the information that the classifier needs to assign an unknown feature vector to one of the available classes.

Parameter Features
The input parameter Features contains the feature vector of the object that is to be classified. The fea-
ture vector must consist of the same features as the feature vectors used for the training samples within
add_sample_class_svm.

Parameter Num
The input parameter Num specifies the number of best classes to be searched for. Generally, Num is set to 1 if only
the class with the best probability is searched for, and to 2 if the second best class is also of interest, e.g., because
the classes overlap. If Mode was set to ’novelty-detection’ in create_class_svm, Num must be set to 1.

Parameter Class
The output parameter Class returns the result of classifying the feature vector with the trained SVM classifier. This
result depends on the Mode that was selected in create_class_svm. If Mode was set to ’one-versus-one’, it
contains the classes ordered by the number of votes of the sub-classifiers. That is, the first element of the returned
tuple is the class with the most votes, the second is the class with the second most votes etc. If Mode was set to
’one-versus-all’, it contains the classes ordered by the value of each sub-classifier. That is, the first element of
the returned tuple is the class with the highest value, the second element is the class with the second best value. If

General Features
Mode was set to ’novelty-detection’, a single value is returned (Num must be set to 1). In particular, the value
is 1 if the feature vector belongs to the trained class and 0 if the feature vector belongs to the rejection class.

5.5 Parameter Setting for GMM

This section goes deeper into the parameter adjustment for a GMM classification. We recommend to first adjust
the parameters so that the classification result is satisfying. The most important parameters that have to be adjusted
to get the GMM classifier work optimally are:

• NumDim (create_class_gmm)
• NumCenters (create_class_gmm)
• CovarType (create_class_gmm)
• ClassPriors (train_class_gmm)

If the classification generally works, you can start to tune the speed. The most important parameters to enhance
the speed are:

• CovarType (create_class_gmm)
• Preprocessing / NumComponents (create_class_gmm)

In the following, we introduce the parameters of the basic GMM operators. The focus is on the parameters for
which the setting is not immediately obvious or for which it is not immediately obvious how they influence the
classification. These are mainly the parameters needed for the creation and training of the classifier. Further
information about these operators as well as the usage of the operators with obvious parameter settings are provided
in the Reference Manual entries for the individual operators.
D-48 Classification of General Features

5.5.1 Adjusting create_class_gmm

A GMM classifier is created with the operator create_class_gmm. There, several properties of the classifier are
defined that are important for the following classification steps. The returned handle is needed (and modified) in
all following steps. The following parameters can be adjusted:

Parameter NumDim
The input parameter NumDim specifies the dimension of the feature vectors used for the training as well as for the
classification.
Note that GMM works optimally only for a limited number of features! If the result of the classification is not
satisfying, maybe you have used too much features as input. As a rule of thumb, a number of 15 features should
not be exceeded (although some applications work also for larger feature vectors). If your application needs
significantly more features, in many cases an MLP, SVM, or k-NN classification is to be preferred.

Parameter NumClasses
The input parameter NumClasses specifies the number of classes.

Parameter NumCenters
As we learned in section 3.5 on page 20 a GMM class can consist of different Gaussian centers (see also fig-
ure 5.10). The input parameter NumCenters defines the number of Gaussian centers per class. You can specify
this number in different ways. That is, you can either specify a single number of centers, then each class has exactly
this number of class centers. Or you can specify the allowed lower and upper bound for the number of centers.
This can be done either with a single range for all classes or with a range for each class individually. From these
bounds, the optimum of centers is determined with the help of the Minimum Message Length Criterion (MML). In
most cases, it is recommended to specify a range for all classes and to start with a high value as upper bound and
the expected number of centers as lower bound. If the classification is successful, you can try to reduce the range
to enhance the speed.

Feature 1: Feature 1:
xx x

xx x
x x

x x
x xx
x xx

x xx

x xx
x

x
x x

x x
x

x
xx x

xx x
xx x

xx x
x

x
x x

xx

x x

xx

NumCenters = 2 NumCenters = 1
x

Feature 2: Feature 2:

Figure 5.10: Number of Gaussian centers of a class: (left) 2 and (right) 1.

Note that if the training is canceled with the error message 3335 (’Internal error while training the GMM’), most
probably the value for NumCenters is not optimal.

Parameter CovarType
The input parameter CovarType defines the type of the covariance matrix used to calculate the probabilities. With
this, you can further constrain the MML, which is used to determine the optimum of centers. Three types of
covariance matrices are available. If you use the default, ’spherical’, the covariance matrix is a scalar multiple
of the identity matrix. With the value ’diag’ a diagonal matrix is obtained and with ’full’ the covariance
matrix is positive definite (see figure 5.11). Note that the flexibility of the centers but also the complexity of the
calculations increases from ’spherical’ over ’diag’ to ’full’. That is, you have to decide whether you want
to increase the flexibility of the classifier or if you want to increase the speed.
 5.5 Parameter Setting for GMM D-49

Feature 1: Feature 1: Feature 1:

x xx x

x xx x

x xx x
x xx

x xx

x xx

x
x

x
x xx

x xx

x xx
x x

x x

x x
x

x
xx x

xx x

xx x
’spherical’ ’diag’ ’full’
x x

x x

x x
xx

xx

xx
xx

xx

xx
x

xx

xx

xx
Feature 2: Feature 2: Feature 2:

Figure 5.11: The covariance type set to (from left to right) ’spherical’, ’diag’, and ’full’.

Parameters Preprocessing / NumComponents

The input parameter Preprocessing defines the type of preprocessing applied to the feature vector for the training
as well as later for the classification or evaluation. A preprocessing of the feature vector can be used to speed up
the training as well as the classification. Sometimes, even the recognition rate can be enhanced.
Available values are ’none’, ’normalization’, ’principal_components’, and ’canonical_variates’.
In most cases, Preprocessing should be set to ’normalization’ as it enhances the speed without loosing
relevant information compared to using no preprocessing (’none’). The values ’principal_components’ and
in rare cases ’canonical_variates’ can be used to enhance the speed. As the preprocessing types are the same
as used for an MLP classification, we refer to section 5.3.1 on page 38 for further information.
The input parameter NumComponents defines the number of components to which the feature vector is reduced
if a preprocessing is selected that reduces the dimension of the feature vector. In particular, NumComponents
has to be adjusted only if Preprocessing is set to ’principal_components’ or ’canonical_variates’.
If Preprocessing is set to ’principal_components’ or ’canonical_variates’ you can use the operator
get_prep_info_class_gmm to determine the optimum number of components as described in section 5.3.1 on

General Features
page 38.

Parameter RandSeed
The coordinates of the centers are initialized by a random number. For the sake of reproducibility the seed value
for this random number is stored in the input parameter RandSeed.

Parameter GMMHandle
The output parameter of create_class_gmm is the GMMHandle, which is needed for all following classification
specific operators.

5.5.2 Adjusting add_sample_class_gmm

A single sample is added to the GMM classifier using add_sample_class_gmm. For the training, several sam-
ples have to be added by successively calling add_sample_class_gmm with different samples. The following
parameters can be adjusted:

Parameter GMMHandle
The input and output parameter GMMHandle is the handle of the classifier that was created with
create_class_gmm and to which the samples subsequently were added with add_sample_class_gmm. After
applying add_sample_class_gmm for all available samples, the handle is prepared for the actual training of the
classifier.

Parameter Features
The input parameter Features contains the feature vector of a sample to be added to the classifier with
add_sample_class_gmm. This feature vector is a tuple of values. Each value describes a specific numeric feature.
Note that the feature vector must consist of real numbers. If you have integer numbers, you have to transform them
into real numbers. Otherwise, an error message is raised.
D-50 Classification of General Features

Parameter ClassID
The input parameter ClassID contains the ID of the class the feature vector belongs to. The ID is an integer
number between 0 and ’Number of Classes -1’. If you created a tuple with class names, the class ID is the
index of the corresponding class name in the tuple.

Parameter Randomize
The input parameter Randomize defines the standard deviation of the Gaussian noise that is added to the training
data. This value is needed mainly for originally integer feature values. There, the modeled Gaussians may be
aligned along axis directions and thus lead to an unusually high number of centers returned by train_class_gmm
(see figure 5.12). This effect can be prevented by setting Randomize to a value larger than 0. According to
experience, a value between 1.5 and 2 in most cases leads to a satisfying result. If the feature vector has been
created from integer data by scaling, Randomize must be scaled with the same scale factor that was used to scale
the original data.
Feature 1: Feature 1:
x

xx
x

xx
x

xx xx
x
x

xx
xx
x

Feature 2: Feature 2:

Integer Feature Vectors

x x

Randomized Feature Vectors

Figure 5.12: Adding noise to integer values: (left) the integer feature vectors lead to many classes, whereas for the
(right) randomized feature vectors one class is obtained.

5.5.3 Adjusting train_class_gmm

The training of the GMM classifier is applied with train_class_gmm. The following parameters can be adjusted:

Parameter GMMHandle
The input and output parameter GMMHandle is the handle of the classifier that was created with
create_class_gmm and for which samples were stored either by adding them via add_sample_class_gmm or by
reading them in with read_samples_class_gmm. After applying train_class_gmm, the handle is prepared for
the actual classification of unknown data. That is, it then contains information about how to separate the classes.

Parameters MaxIter / Threshold

The input parameter MaxIter defines the maximum number of iterations used for the expectation minimization
algorithm. The input parameter Threshold defines the threshold for the relative change of the error for the
expectation minimization algorithm to terminate. By reducing the number of iterations, the speed can be optimized
for specific applications. But note that in most cases, the parameters MaxIter and Threshold should be used with
the default values.

Parameter ClassPriors
The input parameter ClassPriors is used to select the mode for determining the probability of the occurrence
of a class (see also section 3.5 on page 20). That is, ClassPriors determines if a weighting of the classes is
used that is derived from the proportion of the corresponding sample data used for the training (ClassPriors
set to ’training’) or if all classes have the same weight (ClassPriors set to ’uniform’), i.e., the weight is
1/NumClasses for all classes (see figure 5.13). By default, the mode ’training’ is selected, i.e., the probability
of the occurrence of a class is derived from the frequency of the class in the training set. If your training data is not
representative for the frequency of the individual classes, you should use ’uniform’ instead.
 5.5 Parameter Setting for GMM D-51

Class 1 Class 2 Class 1 Class 2

’uniform’ ’training’

Figure 5.13: Probability of the occurrence of a class set to (left) ’uniform’ and (right) ’training’.

Parameter Regularize
The input parameter Regularize is used to prevent the covariance matrix from a collapse which can occur for
linearly dependent data. Here, we recommend to use the default value, which is 0.0001.

Parameter Centers
The output parameter Centers returns the number of found centers per class.

Parameter Iter
The output parameter Iter returns the number of iterations that were executed for the expectation minimization
algorithm for each class.

5.5.4 Adjusting evaluate_class_gmm

General Features
The operator evaluate_class_gmm can be used to evaluate the probabilities for a feature vector to belong to each
of the available classes. If only the probabilities for the most probable classes are searched for, no evaluation is
necessary, as these probabilities are returned also for the final classification of the feature vector. The following
parameters can be adjusted:

Parameter GMMHandle
The input parameter GMMHandle is the handle of the classifier that was previously trained with the operator
train_class_gmm.

Parameter Features
The input parameter Features contains the feature vector that is evaluated. The feature vector must consist of the
same features as the feature vectors used for the training samples within add_sample_class_gmm.

Parameter ClassProb
The output parameter ClassProb returns the a-posteriori probabilities of the given feature vector to belong to each
of the classes.

Parameter Density
The output parameter Density returns the probability density of the feature vector.

Parameter KSigmaProb
The output parameter KSigmaProb describes the probability that another sample lies farther away from the mean.
This value can be used for novelty detection. Then, all feature vectors with a KSigmaProb value below a certain
k-sigma probability, e.g., 0.0001, can be rejected.
D-52 Classification of General Features

5.5.5 Adjusting classify_class_gmm

The operator classify_class_gmm is used to decide to which of the trained classes an unknown feature vector
belongs. The following parameters can be adjusted:

Parameter GMMHandle
The input parameter GMMHandle describes the handle that was created with create_class_gmm, to which samples
were added with add_sample_class_gmm, and that was trained with train_class_gmm. The handle contains
all the information that the classifier needs to assign an unknown feature vector to one of the available classes.

Parameter Features
The input parameter Features contains the feature vector of the object that is to be classified. The fea-
ture vector must consist of the same features as the feature vectors used for the training samples within
add_sample_class_gmm.

Parameter Num
The input parameter Num specifies the number of best classes to be searched for. Generally, Num is set to 1 if only
the class with the best probability is searched for, and to 2 if the second best class is also of interest, e.g., because
the classes overlap.

Parameter ClassID
The output parameter ClassID returns the result of classifying the feature vector with the trained GMM classifier,
i.e., a tuple containing Num elements. That is, if Num is set to 1, a single value is returned that corresponds to the
class with the highest probability. If Num is set to 2, the first element contains the class with the highest probability
and the second element contains the second best class.
The following parameters output the probabilities of the classes. In comparison to the confidence value returned
for an MLP classification (see section 5.3.5 on page 41), the returned values are rather reliable.

Parameter ClassProb
The output parameter ClassProb returns the a-posteriori probabilities of the given feature vector to belong to
each of the classes. In contrast to the ClassProb returned by evaluate_class_gmm, the probability is further
normalized.

Parameter Density
The output parameter Density returns the probability density of the feature vector.

Parameter KSigmaProb
The output parameter KSigmaProb describes the probability that another sample lies farther away from the mean.
This value can be used for novelty detection. Then, all feature vectors with a KSigmaProb value below a certain
k-sigma probability, e.g., 0.0001, can be rejected.

5.6 Parameter Setting for k-NN

This section goes deeper into the parameter adjustment for a k-NN classification. We recommend to first adjust the
parameters so that the classification result is satisfying. The most important parameters that have to be adjusted to
get the k-NN classifier work optimally are:

• NumDim (create_class_knn)
• The kind of results that is returned by classify_class_knn (controlled by the parameter ’method’, which
is set with set_params_class_knn).
 5.6 Parameter Setting for k-NN D-53

If the classification generally works, you can start to tune the speed. The most important parameters to enhance
the speed are:

• The number of neighbors (’k’).

• The maximum number of returned classes (’max_num_classes’).
• The accuracy of the search for nearest neighbors, which is controlled by the parameters ’num_checks’ and
’epsilon’.

All these parameters can be set with the operator set_params_class_knn.

In the following, we introduce the parameters of the basic k-NN operators. The focus is on the parameters for
which the setting is not immediately obvious or for which it is not immediately obvious how they influence the
classification. These are mainly the parameters needed for controlling the kind of result that is returned as well
as the parameters that are used to speed up the classification. Further information about these operators as well
as the usage of the operators with obvious parameter settings is provided in the Reference Manual entries for the
individual operators.

5.6.1 Adjusting create_class_knn

A k-NN classifier is created with the operator create_class_knn. The returned handle is needed in all following
steps. The only parameter that can be set is

Parameter NumDim
The input parameter NumDim specifies the dimension of the feature vectors used for the training as well as for the
classification.

General Features
5.6.2 Adjusting add_sample_class_knn

A single sample is added to the k-NN classifier using add_sample_class_knn. For the training, several sam-
ples have to be added by successively calling add_sample_class_knn with different samples. The following
parameters can be adjusted:

Parameter KNNHandle
The input and output parameter KNNHandle is the handle of the classifier that was created with
create_class_knn and to which the samples subsequently were added with add_sample_class_knn. After
applying add_sample_class_knn for all available samples, the handle is prepared for the actual training of the
classifier.

Parameter Features
The input parameter Features contains the feature vector of a sample to be added to the classifier with
add_sample_class_knn. This feature vector is a tuple of values. Each value describes a specific numeric feature.

Parameter ClassID
The input parameter ClassID contains the ID of the class the feature vector belongs to. The ID is an integer
number between 0 and ’Number of Classes -1’. If you have created a tuple with class names, the class ID is
the index of the corresponding class name in the tuple.

5.6.3 Adjusting train_class_knn

The training of the k-NN classifier is applied with train_class_knn. The following parameters can be adjusted:
D-54 Classification of General Features

Parameter KNNHandle
The input and output parameter KNNHandle is the handle of the classifier that was created with
create_class_knn and for which samples were stored by adding them via add_sample_class_knn. After
applying train_class_knn, the handle is prepared for the actual classification of unknown data, i.e., the internal
representation of the samples is optimized for an efficient search.

Generic Parameter ’num_trees’

This parameter influences the internal representation of the samples and thus the accuracy of the k-NN classification
as well as its runtime. The default value for ’num_trees’ is 4. To speed up the classification, the number of trees
must be set to a lower value. To achieve a more accurate classification result, ’num_trees’ must be set to a higher
value.

Generic Parameter ’normalization’

If ’normalization’ is set to ’true’, the feature vectors are normalized by subtracting the mean of the individual
components of the training vectors and dividing the result by the standard deviation of the individual components
of the training vectors. Hence, the normalized feature vectors have a mean of 0 and a standard deviation of 1. The
normalization does not change the length of the feature vector.
Note that the training samples stored in the k-NN classifier are modified if train_class_knn is called with
’normalization’ set to ’true’, but the original data can be restored at any time by calling train_class_knn with
’normalization’ set to ’false’. If normalization is used, the operator classify_class_knn interprets the input data
as unnormalized and performs normalization internally as it has been defined in the last call to train_class_knn.
If you know the relation between the dimensions of the different features, it is best to apply the normalization
explicitly. For example, if the first feature is given in ’mm’ and the second feature is given in ’m’, scaling the first
with 0.001 (or the second with 1000.0) will typically produce better classification results than using the built-in
normalization of train_class_knn.

5.6.4 Adjusting set_params_class_knn

With set_params_class_knn, some parameters can be set that control the behavior of classify_class_knn.

Generic Parameter ’k’

The parameter ’k’ defines the number of nearest neighbors that are determined during the classification. The
selection of a suitable value for ’k’ depends heavily on the classification task. Generally, larger values of ’k’ lead
to higher robustness against noise, smooth the boundaries between the classes, and lead to longer runtimes during
the classification.
In practice, the best way of finding a suitable value for ’k’ is indeed to try different values and to select the value
for ’k’ that yields the best classification results under the constraint of an acceptable runtime.

Generic Parameters ’method’ and ’max_num_classes’

The parameter ’method’ controls the kind of result that is returned by classify_class_knn while the param-
eter ’max_num_classes’ controls how many different classes may be returned. Note that ’max_num_classes’
is an upper bound for the number of returned classes because the ’k’ nearest neighbors may contain less than
’max_num_classes’ classes.
The default value for ’max_num_classes’ is 1. In this case, only the best rated class is returned, which is often
sufficient. If you need information about the reliability of the classification result, ’max_num_classes’ should be
set to a value larger than 1 and the ratings of the returned classes, which are given in the output parameter Rating
of the operator classify_class_knn, should be analyzed. For example, you can check if the rating of the first
returned class is significantly better than that of the second one. If this is not the case, the classification of this
specific sample is not reliable. In this case, application knowledge may help to decide, which of the best rated
results is the correct one.
If ’method’ is set to ’classes_distance’, classify_class_knn returns each class that exists in the set of the k-
nearest neighbors. For each returned class, its smallest distance to the sample to be classified is returned. The
 5.6 Parameter Setting for k-NN D-55

returned classes are sorted according to this distance, i.e., the first element of the returned classes contains the
nearest neighbor. ’classes_distance’ is the default value for the ’method’.
If ’method’ is set to ’classes_frequency’, classify_class_knn performs a simple majority vote (see section 3.6
on page 22) and returns those classes that occur among the ’k’ nearest neighbors sorted according to their relative
frequency. For example, if ’k’ is set to 10 and among the 10 nearest neighbors, there are 7 samples that belong to
class 9 and 3 samples that belong to class 4, the tuple [9, 4] is returned.
If ’method’ is set to ’classes_weighted_frequencies’, classify_class_knn performs a weighted majority vote
(see section 3.6 on page 22) and returns those classes that occur among the ’k’ nearest neighbors sorted according
to their relative frequency weighted with the distances of the individual neighbors from the sample to be classified.
Thus, classes are better rated if they provide neighbors very close to the sample to be classified.
If ’method’ is set to ’neighbors_distance’, classify_class_knn returns the indices of the ’k’ nearest neighbors
and their distances. This may be useful if none of the above described options fits your requirements. In this
case, you can use the information about the nearest neighbors and their distances to the sample to be classified to
implement your own voting scheme.

Generic Parameters ’num_checks’ and ’epsilon’

In order to provide a really fast k-NN classifier, HALCON uses an approximate algorithm for the search for the
’k’ nearest neighbors. The two parameters ’num_checks’ and ’epsilon’ allow to control the trade-off between
speed and accuracy of this search. Typically, adjusting the parameter ’num_checks’ has a greater effect than
adjusting the parameter ’epsilon’.
’num_checks’ sets the maximum number of runs through the internal search trees. The default value is 32. To
speed up the search, use a lower value (that is greater than 0). To perform an exact search, ’num_checks’ must
be set to 0.
’epsilon’ sets a stop criterion for the search. The default value is 0.0. To potentially speed up the search, use a
higher value.

General Features
5.6.5 Adjusting classify_class_knn

The operator classify_class_knn is used to decide to which of the trained classes an unknown feature vector
belongs. The following parameters can be adjusted:

Parameter KNNHandle
The input parameter KNNHandle describes the handle that was created with create_class_knn, to which samples
were added with add_sample_class_knn, and that was trained with train_class_knn. The handle contains
all the information that the classifier needs to assign an unknown feature vector to one of the available classes.

Parameter Features
The input parameter Features contains the feature vector of the object that is to be classified. The feature
vector must consist of the same kind of features as the feature vectors used for the training samples within
add_sample_class_knn.

Parameter Result
The output parameter Result returns the result of classifying the feature vector with the trained k-NN classifier.
This result depends on the ’method’ that was selected in set_params_class_knn (see section 5.6.4 on page 54
above for a detailed description of set_params_class_knn).

Parameter Rating
The output parameter Rating returns the distances of the returned classes from their nearest neighbor(s). This
result depends on the ’method’ that was selected in set_params_class_knn (see section 5.6.4 on page 54 for the
description of set_params_class_knn).
D-56 Classification of General Features
 Classification for Image Segmentation D-57

Chapter 6

Classification for Image Segmentation

If classification is used to find objects in an image, the individual pixels of an image are classified according to the
features ’color’ or ’texture’ and all pixels belonging to the same class are combined in a region representing the
desired object. That is, the image is segmented into regions of different classes.
For image segmentation, the pixels of an image are classified according to a set of available classes, which are
defined by the training samples. The training samples are image regions of known classes. The features used for
the training of the classes as well as for the classification are color or texture. Such a pixel-based classification can
be realized by several approaches in HALCON. These are MLP, SVM, GMM, and k-NN classifiers (see section 6.1)
and some simple but fast classifiers that use a 2D histogram to segment two-channel images (see section 6.2 on
page 72) or that apply a hyperbox or Euclidean classification for multi-channel images (see section 6.3 on page
73).

6.1 Approach for MLP, SVM, GMM, and k-NN

The set of operators used for image segmentation with MLP, SVM, GMM, and k-NN in parts corresponds to the
set of operators used for the general classification described in section 5 on page 31. Operators that are specific for
image segmentation are those that add the training samples to the classifier and the operators used for the actual
classification. These are used instead of the corresponding general operators.
In section 6.1.1, the general approach for image segmentation is illustrated by different examples that on the one

Segmentation
hand show how to segment different citrus fruits from the background using color and on the other hand show how
to apply novelty detection for a regular mesh using texture. In section 6.1.2 the steps of an image segmentation
and the involved operators are listed for a brief overview. The parameters of the operators that are specific for
image segmentation are introduced in more detail in section 6.1.3 for MLP, section 6.1.4 for SVM, section 6.1.5
for GMM, and section 6.1.6 for k-NN.
Finally, section 6.1.7 shows how to speed up image segmentation for images with a maximum of three image
channels by applying a classification that is based on look-up tables (LUT). Here, the trained classifier is used to
create a look-up table that stores every possible response of the MLP, SVM, GMM, or k-NN classifier, respectively.
Using the LUT-accelerated classifier instead of the original trained classifier, the class of every image point can
be taken directly from the LUT instead of being calculated expensively. But note that for a LUT-accelerated
classification also additional memory is needed and the runtime for the offline part is increasing.

6.1.1 General Approach

Figure 6.1 shows the general approach for image segmentation. The main difference to the general classification
approach is that on the one hand the objects to classify are restricted to pixels and on the other hand the features
are not explicitly extracted but automatically derived from the different channels of a color or texture image. Thus,
you do not have to apply a feature extraction for the training and once again for the classification, but simply apply
a training using some sample regions of multi-channel images. Then you can immediately use the trained classifier
to segment images into regions of the trained color or texture classes.
D-58 Classification for Image Segmentation

Create Classifier

Investigate Known Regions of Multi−Channel Images

For each
Use Color Image or Create Texture Image −> Sample Image

Image
Specify Regions of Known Color or Texture Class −> Sample Classes
Add Sample Image and Sample Classes to Classifier

Train Classifier

Investigate Unknown Image

Segment Image into Regions of Different Classes (Classification)

Figure 6.1: The basic steps of image segmentation.

Besides the classical image segmentation, the operators for SVM, GMM, and k-NN can be used also for novelty
detection. In this case, for SVM only one class is trained and the classification returns the pixels that significantly
deviate from the class. For GMM and k-NN, one or more classes can be trained and the classification rejects the
pixels that significantly deviate from any of the classes. That is, the “novelties” (or defects, respectively) within
an image are those pixels that are not assigned to one of the trained classes. The following examples demonstrate
how to apply classical image segmentation and novelty detection.

[Link] Image Segmentation

The example %HALCONEXAMPLES%\solution_guide\classification\segment_citrus_fruits.hdev

shows how to segment an image to separate lemons and oranges from their background. The lemons and oranges
were already introduced in section 3 on page 15. There, a general classification with shape features was proposed
to distinguish between lemons and oranges. The corresponding example can be found in section 8.3.1 on page 96.
Here, the lemons and oranges are separated from the background using their color, i.e., the feature vectors are built
by the gray values of three image channels. The classification is applied with the MLP approach. The operator
names for the GMM, SVM, and k-NN classification differ only in their ending. That is, if you want to apply a
GMM, SVM, or k-NN classification, you mainly have to replace ’mlp’ by ’gmm’, ’svm’, or ’knn’, respectively in
the specific operator names and adjust different parameters. The parameters of the operators that correspond to
the general classification and their selection are described in section 5.3 for MLP, section 5.4 for SVM, section 5.5
for GMM, and section 5.6 for k-NN. The parameters of the operators that are specific for image segmentation
and their selection are described in section 6.1.3 for MLP, section 6.1.4 for SVM, section 6.1.5 for GMM, and
section 6.1.6 for k-NN.
The program starts with the creation of an MLP classifier. Following the instructions given in section 5.3.1 on
page 37 the parameter NumInput is set to 3 as the images consist of three channels, which leads to three features
for the feature vectors, the parameter NumHidden is set to 3 so that it is in a similar value range as NumInput and
NumOutput. NumOutput is set to 3 as three classes are used: one for the oranges, one for the lemons, and one for
the background. OutputFunction must be set to ’softmax’ for image segmentation. Preprocessing is set to
’normalization’, so NumComponents can be ignored. The operator returns the handle of the new classifier that
is needed for the following steps.

create_class_mlp (3, 3, 3, 'softmax', 'normalization', 10, 42, MLPHandle)

Now, an image containing oranges is read and a region for the class ’orange’ and one for the class ’background’
are created (see figure 6.2, left). As no lemon is contained in the image, an empty region is created by
gen_empty_region. All three regions are concatenated to a tuple with concat_obj and are then added together
with the input image and the handle of the classifier to the classifier with add_samples_image_class_mlp.
 6.1 Approach for MLP, SVM, GMM, and k-NN D-59

read_image (Image, 'color/citrus_fruits_01')

gen_rectangle1 (OrangeRegion, 100, 130, 230, 200)
gen_rectangle1 (BackgroundRegion, 30, 20, 50, 50)
gen_empty_region (EmptyRegion)
gen_empty_obj (TrainingRegions1)
concat_obj (TrainingRegions1, OrangeRegion, TrainingRegions1)
concat_obj (TrainingRegions1, EmptyRegion, TrainingRegions1)
concat_obj (TrainingRegions1, BackgroundRegion, TrainingRegions1)
add_samples_image_class_mlp (Image, TrainingRegions1, MLPHandle)

A second image is read that contains lemons. Now, a region for the class ’lemons’ and one for the class
’background’ are generated (see figure 6.2, right) and are concatenated together with an empty region to
a tuple of regions. The sequence of the contained regions is the same as for the image with the oranges,
i.e., the first element contains the region for oranges (in this case an empty region), the second element con-
tains the region for lemons, and the third element contains the region for the background. Then, the operator
add_samples_image_class_mlp is called again to extent the samples that are already added to the classifier.

Figure 6.2: Regions from two images are used as training regions.

read_image (Image, 'color/citrus_fruits_03')

gen_rectangle1 (LemonRegion, 180, 130, 230, 240)
gen_rectangle1 (BackgroundRegion, 400, 20, 430, 50)

Segmentation
gen_empty_obj (TrainingRegions2)
concat_obj (TrainingRegions2, EmptyRegion, TrainingRegions2)
concat_obj (TrainingRegions2, LemonRegion, TrainingRegions2)
concat_obj (TrainingRegions2, BackgroundRegion, TrainingRegions2)
add_samples_image_class_mlp (Image, TrainingRegions2, MLPHandle)

After adding all training regions, the classifier is trained.

train_class_mlp (MLPHandle, 200, 1, 0.01, Error, ErrorLog)

Now, a set of images is read and segmented according to the rules that the classifier derived from the training. The
result is a region for each class.
D-60 Classification for Image Segmentation

for I := 1 to 15 by 1
read_image (Image, 'color/citrus_fruits_' + I$'.2d')
classify_image_class_mlp (Image, ClassRegions, MLPHandle, 0.5)
select_obj (ClassRegions, ClassOranges, 1)
select_obj (ClassRegions, ClassLemons, 2)
select_obj (ClassRegions, ClassBackground, 3)
dev_set_draw ('fill')
dev_display (Image)
dev_set_color ('slate blue')
dev_display (ClassBackground)
dev_set_color ('goldenrod')
dev_display (ClassOranges)
dev_set_color ('yellow')
dev_display (ClassLemons)

The result of the segmentation is visualized by three different colors. Note that the colors applied in the example
and the colors used for the representation in figure 6.3 vary because different colors are suited for print purposes and
for the presentation on a screen. Note further, that the erroneously classified pixels are caused by the overlapping
classes that occur because of gray shadings that affected both types of fruits.

Figure 6.3: Segmentation of the image into the classes (dim gray) ’background’, (gray) ’oranges’, and (white)
’lemons’. Erroneously classified pixels at the border of the lemons occur, because the shadows at the
border of both fruit types have the same color.

For a better result, the illumination could have been adjusted more carefully to avoid shadings and reflections of the
fruits. Additionally, many more samples would have been required for a ’real’ application. Note that this example
mainly aims to demonstrate the general approach of an image segmentation. But even with the few erroneously
classified pixels, a class decision can be found for each fruit. For that, we apply a post processing. That is, for
each fruit class we use morphological operators to close small gaps, apply the operator connection to separate
connected components, and then select those shapes from the connected components that exceed a specific size.
Additionally, holes inside the regions are filled up with fill_up and the shapes of the regions are transformed to
their convex hulls with shape_trans. The result is shown in figure 6.4.
 6.1 Approach for MLP, SVM, GMM, and k-NN D-61

closing_circle (ClassOranges, RegionClosingOranges, 3.5)

connection (RegionClosingOranges, ConnectedRegionsOranges)
select_shape (ConnectedRegionsOranges, SelectedRegionsOranges, 'area', \
'and', 20000, 99999)
fill_up (SelectedRegionsOranges, RegionFillUpOranges)
shape_trans (RegionFillUpOranges, RegionFillUpOranges, 'convex')
closing_circle (ClassLemons, RegionClosingLemons, 3.5)
connection (RegionClosingLemons, ConnectedRegionsLemons)
select_shape (ConnectedRegionsLemons, SelectedRegionsLemons, 'area', \
'and', 15000, 99999)
fill_up (SelectedRegionsLemons, RegionFillUpLemons)
shape_trans (RegionFillUpLemons, RegionFillUpLemons, 'convex')
dev_display (Image)
dev_set_draw ('margin')
dev_set_color ('goldenrod')
dev_display (RegionFillUpOranges)
dev_set_color ('yellow')
dev_display (RegionFillUpLemons)
endfor

Figure 6.4: Segmentation result after postprocessing.

Segmentation
For the images at hand, alternatively also a general classification using shape features can be applied like described
in section 8.3.1 on page 96.

[Link] Novelty Detection with SVM

The example program %HALCONEXAMPLES%\hdevelop\Segmentation\Classification\

novelty_detection_svm.hdev shows how to apply a novelty detection with SVM. For a novelty detection
with SVM, a single class is trained and the classification is used to find all regions of an image that do not belong
to this class.
The program trains the texture of a regular plastic mesh. Before creating a classifier, a rectangle is generated that
is used later as region of interest. This region of interest is necessary, because the images of the plastic mesh to be
inspected do not contain an integer number of mesh cells. Thus, if the original image would have been trained and
classified, the texture filters that are applied to create a multi-channel texture image would probably return artifacts
at the image borders.

gen_rectangle1 (Rectangle, 10, 10, Height / 2 - 11, Width / 2 - 11)

The SVM classifier is created with create_class_svm. With SVM, novelty detection is a two-class problem
(see also section 5.4.1 on page 44) and the classifier must be explicitly set to ’novelty-detection’ using the
parameter Mode. Additionally, the KernelType must be set to ’rbf’.
D-62 Classification for Image Segmentation

create_class_svm (5, 'rbf', 0.01, 0.0005, 1, 'novelty-detection', \

'normalization', 5, SVMHandle)

Then, all training images, i.e., images containing a good mesh, are read in a loop and scaled down by a factor of
two. This is done, so that the textures can be optimally filtered with a filter size of 5x5 when creating a multi-
channel texture image within the procedure gen_texture_image. Theoretically, also the original size could be
used with a filter size of 10x10, but this would need too much time and the accuracy for the smaller image is
sufficient for the application. The procedure gen_texture_image creates a multi-channel texture image for each
image. This is then passed together with the region of interest to the operator add_samples_image_class_svm
to add the sample region to the classifier.

for J := 1 to 5 by 1
read_image (Image, 'plastic_mesh/plastic_mesh_' + J$'02')
zoom_image_factor (Image, ImageZoomed, 0.5, 0.5, 'constant')
disp_message (WindowHandle, 'Adding training samples...', 'window', 12, \
12, 'black', 'true')
gen_texture_image (ImageZoomed, ImageTexture)
add_samples_image_class_svm (ImageTexture, Rectangle, SVMHandle)
endfor

The procedure gen_texture_image creates the multi-channel image by applying the texture filter texture_laws
to the zoomed image with varying parameters and combining the differently filtered images to one image using
compose5. The result is additionally smoothed before the procedure returns the final texture image.

texture_laws (Image, ImageEL, 'el', 5, 5)

texture_laws (Image, ImageLE, 'le', 5, 5)
texture_laws (Image, ImageES, 'es', 1, 5)
texture_laws (Image, ImageSE, 'se', 1, 5)
texture_laws (Image, ImageEE, 'ee', 2, 5)
compose5 (ImageEL, ImageLE, ImageES, ImageSE, ImageEE, ImageLaws)
smooth_image (ImageLaws, ImageTexture, 'gauss', 5)

After adding all training regions to the classifier, the classifier is trained with train_class_svm. To speed up the
classification, the resulting support vectors are reduced with reduce_class_svm, which leads to the new classifier
SVMHandleReduced. Note, that this operator is specific for SVM.

train_class_svm (SVMHandle, 0.001, 'default')

reduce_class_svm (SVMHandle, 'bottom_up', 2, 0.001, SVMHandleReduced)

Now, the novelty detection is applied to several images. That is, each image is transformed to a multi-channel
texture image like described for the training part. This time, as classify_class_svm needs only the image and
not a region as input, the image is additionally reduced to the domain of the region of interest. The reduced image
is then passed to the operator classify_class_svm for novelty detection. The output parameter ClassRegions
(here called Errors) returns all pixels of the image that do not belong to the trained texture. With a set of
morphological operators and a blob analysis it is checked if connected components that exceed a certain size exist,
i.e., if the image contains significant ’novelties’.
 6.1 Approach for MLP, SVM, GMM, and k-NN D-63

for J := 1 to 14 by 1
read_image (Image, 'plastic_mesh/plastic_mesh_' + J$'02')
zoom_image_factor (Image, ImageZoomed, 0.5, 0.5, 'constant')
gen_texture_image (ImageZoomed, ImageTexture)
reduce_domain (ImageTexture, Rectangle, ImageTextureReduced)
classify_image_class_svm (ImageTextureReduced, Errors, SVMHandleReduced)
opening_circle (Errors, ErrorsOpening, 3.5)
closing_circle (ErrorsOpening, ErrorsClosing, 10.5)
connection (ErrorsClosing, ErrorsConnected)
select_shape (ErrorsConnected, FinalErrors, 'area', 'and', 300, 1000000)
count_obj (FinalErrors, NumErrors)
if (NumErrors > 0)
disp_message (WindowHandle, 'Mesh not OK', 'window', 12, 12, 'red', \
'true')
else
disp_message (WindowHandle, 'Mesh OK', 'window', 12, 12, \
'forest green', 'true')
endif
endfor

[Link] Novelty Detection with GMM or k-NN

The example program %HALCONEXAMPLES%\hdevelop\Segmentation\Classification\

novelty_detection_gmm.hdev shows how to apply a novelty detection with GMM. Generally,
the example does the same as %HALCONEXAMPLES%\hdevelop\Segmentation\Classification\
novelty_detection_svm.hdev did. That is, the same images are used and the results are rather similar
(see figure 6.5). The significant differences between novelty detection with GMM and with SVM concern the
parameter settings when creating the classifier and the output returned when classifying an image. The novelty
detection is presented here with the GMM approach. The operator names for the GMM and k-NN classification
differ only in their ending. That is, if you want to apply a k-NN classification for the novelty detection, you mainly
have to replace ’gmm’ with ’knn’ in the specific operator names and adjust some parameters.
When creating the classifier for GMM, in contrast to SVM, no explicit parameter for novelty detection is needed.
Here, simply a classifier for a single class (NumClasses set to 1) is created.

create_class_gmm (5, 1, [1, 5], 'spherical', 'normalization', 5, 42, \

GMMHandle)

Segmentation
When classifying an image, the output parameter ClassRegions (here called Correct) of the operator
classify_image_class_gmm, in contrast to the novelty detection with SVM, does not return a region built by
erroneous pixels but a region that is built by pixels that belong to the trained texture class. To obtain the erroneous
region, the difference between the input region and the returned region has to be calculated using difference.

classify_image_class_gmm (ImageTextureReduced, Correct, GMMHandle, \

0.000002)
difference (Rectangle, Correct, Errors)

Both significant differences occur because the GMM classifier by default returns only those regions that precisely
belong to the trained classes (within a specified threshold) and thus automatically rejects all other pixels. For SVM,
parts that do not belong to a class can only be determined for two-class problems, i.e., when explicitly setting the
Mode to ’novelty-detection’. Otherwise, SVM assigns all pixels to the available classes, even if some pixels
do not significantly match any of them. When explicitly setting a parameter for novelty detection, it is obvious
that the returned region should show the requested novelties. For GMM, no specific parameter has to be set, i.e.,
the novelty detection is realized by applying a regular image segmentation. Thus, the returned region shows the
parts of the image that match the trained class. Note that for the GMM classifier novelty detection is not restricted
to a two-class problem and image segmentation, but can be applied also for multi-class problems and general
classification (see section 5.5.5 on page 52).
D-64 Classification for Image Segmentation

Novelty−Detection with SVM Novelty−Detection with GMM

Figure 6.5: Novelty detection is used to extract parts of an image or region that do not fit to a trained pattern and
can be applied with SVM or GMM.

6.1.2 Involved Operators (Overview)

This section gives a brief overview on the operators that are provided for MLP, SVM, GMM, and k-NN classifica-
tion for image segmentation. In particular, first the operators for the basic steps and then the advanced operators
used for image segmentation are introduced.

[Link] Basic Operators

Summarizing the information obtained in section 6.1.1 on page 57, the image segmentation consists of the follow-
ing basic steps and operators, which are applied in the same order as listed here. Note that the steps are similar to
the steps applied for a general classification (see section 5.2.1 on page 34), mainly the step for adding the samples
and the step for the actual classification vary.

1. Create a classifier. Here, some important properties of the classifier are defined. The returned handle is
needed in all later classification steps. Each classification step modifies this handle. This step corresponds
to the approach of the general classification.
• create_class_mlp
• create_class_svm
• create_class_gmm
• create_class_knn
2. Predefine the sequence in which the classes are defined and later accessed, i.e., define the correspondences
between the class IDs and the class names, respectively define the colors that visualize the different classes.
This step may as well be applied before the creation of the classifier.
3. Add samples, i.e., a region for each class to the classifier. In contrast to the general classification, a single
operator call can be used to add all sample regions at once. The sequence of the added regions define the
classes, i.e., the first region is class 0, the second is class 1 etc. Having several images for the training, the
operator can be called multiple times, then the sample regions for the classes must be defined in the same
sequence. If one of the classes is not represented in an image, an empty region has to be passed. This step
significantly differs from the approach of the general classification.
• add_samples_image_class_mlp
• add_samples_image_class_svm
• add_samples_image_class_gmm
• add_samples_image_class_knn
 6.1 Approach for MLP, SVM, GMM, and k-NN D-65

4. Train the classifier, i.e., use the added samples to obtain the boundaries between the classes. This step
corresponds to the approach of the general classification.
• train_class_mlp
• train_class_svm
• train_class_gmm
• train_class_knn
5. Store the used samples to file and access them in a later step (optionally). This step corresponds to the
approach of the general classification.
• write_samples_class_mlp and read_samples_class_mlp
• write_samples_class_svm and read_samples_class_svm
• write_samples_class_gmm and read_samples_class_gmm
• Note that there are no operators for writing and reading the samples of a k-NN classifier separately
because the samples are an intrinsic component of the k-NN classifier. Use write_class_knn and
read_class_knn instead.
6. Store the trained classifier to file and read it from file again. This step corresponds to the approach of the
general classification.
• write_class_mlp (default file extension: .gmc) and read_class_mlp
• write_class_svm (default file extension: .gsc) and read_class_svm
• write_class_gmm (default file extension: .ggc) and read_class_gmm
• write_class_knn (default file extension: .gnc) and read_class_knn
Note that the samples cannot be deleted from a k-NN classifier because they are an intrinsic component
of this classifier.
7. Segment the image by classification. That is, insert a new image and use one of the following operators to
segment the image into regions of different classes. This step significantly differs from the approach of the
general classification.
• classify_image_class_mlp
• classify_image_class_svm
• classify_image_class_gmm
• classify_image_class_knn

Segmentation
Besides the basic steps of a classification, some additional steps and operators can be applied if suitable. These
advanced operators are similar for image segmentation and general classification.

[Link] Advanced Operators

Especially if the training and classification do not lead to a satisfying result, it is helpful to access some information
that is implicitly contained in the model. Available steps to query information are:

• Access an individual sample from the training data. This is needed, e.g., to check
the correctness of its class assignment. The sample had to be stored previously
by the operator add_samples_image_class_mlp, add_samples_image_class_svm,
add_samples_image_class_gmm, or add_samples_image_class_knn, respectively.
– get_sample_class_mlp
– get_sample_class_svm
– get_sample_class_gmm
– get_sample_class_knn
• Get the number of samples that are stored in the training data. The obtained number is needed, e.g., to access
the individual samples or to know how much individual samples you can access.
D-66 Classification for Image Segmentation

– get_sample_num_class_mlp
– get_sample_num_class_svm
– get_sample_num_class_gmm
– get_sample_num_class_knn
• Get information about the content of the preprocessed feature vectors. This information is reasonable, if the
parameter Preprocessing was set to ’principal_components’ or ’canonical_variates’ during the
creation of the classifier. Then, you can check if the information that is contained in the preprocessed feature
vector still contains significant data or if a different preprocessing parameter, e.g., ’normalization’, is to
be preferred.
– get_prep_info_class_mlp
– get_prep_info_class_svm
– get_prep_info_class_gmm
– Note that this kind of operator is not available for k-NN classifiers, because they do not provide the
above mentioned preprocessing options.
• Get the parameter values that were set during the creation of the classifier. This is needed if the offline
training and the online classification are separated and the information about the training part is not available
anymore.
– get_params_class_mlp
– get_params_class_svm
– get_params_class_gmm
– get_params_class_knn

Furthermore, there are operators that are available only for specific classifiers. In particular,

• For SVM you can reduce the number of support vectors returned by the offline training to speed up the
following online classification.
– reduce_class_svm
• Additionally, for SVM the number or index of the support vectors can be determined after the training. This
is suitable for the visualization of the support vectors and thus for diagnostic reasons.
– get_support_vector_num_class
– get_support_vector_class
• For k-NN classifiers, you can set various parameters with
– set_params_class_knn.
See section 5.6.4 on page 54 for a detailed descriptions of this operator.

If you want to speed up an image segmentation for images with a maximum of three image channels, you can
apply a classification that is based on look-up tables (LUT, see also section 6.1.7 on page 70). Then, the approach
differs from the basic image segmentation as follows:

• After creating a classifier, adding samples to it and training it, the result of the training is stored in a LUT-
accelerated classifier.
– create_class_lut_mlp
– create_class_lut_svm
– create_class_lut_gmm
– create_class_lut_knn
• For the actual image segmentation, instead of classify_image_class_mlp,
classify_image_class_svm, classify_image_class_gmm, or classify_image_class_knn,
respectively, the LUT-accelerated classifier is applied.
– classify_image_class_lut
 6.1 Approach for MLP, SVM, GMM, and k-NN D-67

6.1.3 Parameter Setting for MLP

Most rules for the parameter setting for an image segmentation using MLP classification correspond to the rules for
the general classification with MLP (see section 5.3 on page 37). As the most important operators and parameters
to adjust are listed and described there, here only the parameter settings for the operators that are specific for
image segmentation are described. These are the operators that add the training samples, i.e., sample regions, to
the classifier and those that segment the unknown image.

[Link] Adjusting add_samples_image_class_mlp

For image segmentation with MLP, sample regions are added to the classifier with
add_samples_image_class_mlp. In contrast to the samples used for a general classification, here, no
feature vectors have to be generated explicitly but are implicitly given by the gray values of each pixel of the
respective input region in all channels of the input image. The dimension of the feature vectors depends on the
number of channels of the image containing the sample regions. The quality of the samples is very important for
the quality of the classification. Section 4.3 on page 29 provides hints how to select a set of suitable samples. For
the operator add_samples_image_class_mlp, the following parameters have to be set:

• Image: The image that contains the sample regions

• ClassRegions: The tuple containing the training regions. Here, one region per class is defined. The number
of classes corresponds to the number of regions, and thus, the label of an individual class corresponds to the
position of the training region in the tuple, i.e., its index.
• MLPHandle: The handle of the classifier that was created with create_class_gmm

The operator returns the modified handle of the classifier (MLPHandle).

[Link] Adjusting classify_image_class_mlp

With the operator classify_image_class_mlp a new image is segmented into regions of the classes that were
trained with train_class_mlp using the samples that were added with add_samples_image_class_mlp. The
following parameters have to be set:

• Image: The image that has to be segmented into the classes that were trained with train_class_gmm.
• MLPHandle: The handle of the classifier that was created with create_class_mlp, to which samples were

Segmentation
added with add_samples_image_class_mlp, and that was trained with train_class_mlp.
• RejectionThreshold: The threshold on the probability measure returned by the classification. All pixels
having a probability below RejectionThreshold are not assigned to any class.

The operator returns a tuple of regions in ClassRegions. This tuple contains one region for each class. The
sequence of the classes corresponds to the sequence used for the training regions that were added to the classifier
with add_samples_image_class_mlp.
In contrast to the general classification using classify_class_mlp described in section 5.3.5 on page 41, no
confidence for the classification but a rejection class is returned. This rejection class depends on the selected
rejection threshold. Note that the returned rejection class has not the same quality than the rejection class returned
for a GMM classification, as the MLP classification is typically influenced by outliers (see section 5.3.5 on page
42). Thus, for MLP the explicit training of a rejection class is recommended.

6.1.4 Parameter Setting for SVM

Most rules for the parameter setting for an image segmentation using SVM classification correspond to the rules for
the general classification with SVM (see section 5.4 on page 42). As the most important operators and parameters
to adjust are listed and described there, here only the parameter settings for the operators that are specific for
image segmentation are described. These are the operators that add the training samples, i.e., sample regions, to
the classifier and those that segment the unknown image.
D-68 Classification for Image Segmentation

[Link] Adjusting add_samples_image_class_svm

For image segmentation with SVM, sample regions are added to the classifier with
add_samples_image_class_svm. In contrast to the samples used for a general classification, here, no
feature vectors have to be generated explicitly but are implicitly given by the gray values of each pixel of the
respective input region in all channels of the input image. The dimension of the feature vectors depends on the
number of channels of the image containing the sample regions. The quality of the samples is very important for
the quality of the classification. Section 4.3 on page 29 provides hints how to select a set of suitable samples. For
the operator add_samples_image_class_svm, the following parameters have to be set:

• Image: The image that contains the sample regions

• ClassRegions: The tuple containing the training regions. Here, one region per class is defined. The number
of classes corresponds to the number of regions, and thus, the label of an individual class corresponds to the
position of the training region in the tuple, i.e., its index.
• SVMHandle: The handle of the classifier that was created with create_class_svm

The operator returns the modified handle of the classifier (SVMHandle).

[Link] Adjusting classify_image_class_svm

With the operator classify_image_class_svm a new image is segmented into regions of the classes that were
trained with train_class_svm using the samples that were added with add_samples_image_class_svm. The
following parameters have to be set:

• Image: The image that has to be segmented into the classes that were trained with train_class_svm.
• SVMHandle: The handle of the classifier that was created with create_class_svm, to which samples were
added with add_samples_image_class_svm, and that was trained with train_class_svm.

The operator returns a tuple of regions in ClassRegions. This tuple contains one region for each class. The
sequence of the classes corresponds to the sequence used for the training regions that were added to the classifier
with add_samples_image_class_svm.

6.1.5 Parameter Setting for GMM

Most rules for the parameter setting for an image segmentation using GMM classification correspond to the rules
for the general classification with GMM (see section 5.5 on page 47). As the most important operators and param-
eters to adjust are listed and described there, here only the parameter settings for the operators that are specific for
image segmentation are described. These are the operators that add the training samples, i.e., sample regions, to
the classifier and those that segment the unknown image.

[Link] Adjusting add_samples_image_class_gmm

For image segmentation with GMM, sample regions are added to the classifier with
add_samples_image_class_gmm. In contrast to the samples used for a general classification, here, no
feature vectors have to be generated explicitly but are implicitly given by the gray values of each pixel of the
respective input region in all channels of the input image. The dimension of the feature vectors depends on the
number of channels of the image containing the sample regions. The quality of the samples is very important for
the quality of the classification. Section 4.3 on page 29 provides hints how to select a set of suitable samples. For
the operator add_samples_image_class_gmm, the following parameters have to be set:

• Image: The image that contains the sample regions

• ClassRegions: The tuple containing the training regions. Here, one region per class is defined. The number
of classes corresponds to the number of regions, and thus, the label of an individual class corresponds to the
position of the training region in the tuple, i.e., its index.
 6.1 Approach for MLP, SVM, GMM, and k-NN D-69

• GMMHandle: The handle of the classifier that was created with create_class_gmm
• Randomize: The parameter that handles undesired effects that may occur for originally integer feature values
(see also section 5.5.2 on page 50).

The operator returns the modified handle of the classifier (GMMHandle).

[Link] Adjusting classify_image_class_gmm

With the operator classify_image_class_gmm a new image is segmented into regions of the classes that were
trained with train_class_gmm using the samples that were added with add_samples_image_class_gmm. The
following parameters have to be set:

• Image: The image that has to be segmented into the classes that were trained with train_class_gmm.

• GMMHandle: The handle of the classifier that was created with create_class_gmm, to which samples were
added with add_samples_image_class_gmm, and that was trained with train_class_gmm.
• RejectionThreshold: The threshold on the K-sigma probability (KSigmaProb) measure returned
by the classification (see also section 5.5.5 on page 52). All pixels having a probability below
RejectionThreshold are not assigned to any class.

The operator returns a tuple of regions in ClassRegions. This tuple contains one region for each class. The
sequence of the classes corresponds to the sequence used for the training regions that were added to the classifier
with add_samples_image_class_gmm.
In contrast to the general classification using classify_class_gmm described in section 5.5.5 on page 52, no
probabilities for the classes but a rejection class is returned. This rejection class depends on the selected rejection
threshold.

6.1.6 Parameter Setting for k-NN

Most rules for the parameter setting for an image segmentation using k-NN classification correspond to the rules for
the general classification with k-NN (see section 5.6 on page 52). As the most important operators and parameters
to adjust are listed and described there, here only the parameter settings for the operators that are specific for
image segmentation are described. These are the operators that add the training samples, i.e., sample regions, to

Segmentation
the classifier and those that segment the unknown image.

[Link] Adjusting add_samples_image_class_knn

For image segmentation with k-NN, sample regions are added to the classifier with
add_samples_image_class_knn. In contrast to the samples used for a general classification, here, no
feature vectors have to be generated explicitly but are implicitly given by the gray values of each pixel of the
respective input region in all channels of the input image. The dimension of the feature vectors depends on the
number of channels of the image containing the sample regions. The quality of the samples determines the quality
of the classification. Section 4.3 on page 29 provides hints how to select a set of suitable samples. For the operator
add_samples_image_class_knn, the following parameters have to be set:

• Image: The image that contains the sample regions

• ClassRegions: The tuple containing the training regions. Here, one region per class is defined. The number
of classes corresponds to the number of regions, and thus, the label of an individual class corresponds to the
position of the training region in the tuple, i.e., its index.
• KNNHandle: The handle of the classifier that was created with create_class_knn

The operator returns the modified handle of the classifier (KNNHandle).

D-70 Classification for Image Segmentation

[Link] Adjusting classify_image_class_knn

With the operator classify_image_class_knn a new image is segmented into regions of the classes that were
trained with train_class_knn using the samples that were added with add_samples_image_class_knn. The
following parameters have to be set:

• Image: The image that has to be segmented into the classes that were trained with train_class_knn.
• KNNHandle: The handle of the classifier that was created with create_class_knn, to which samples were
added with add_samples_image_class_knn, and that was trained with train_class_knn.
• RejectionThreshold: The threshold on the distance returned by the classification (see also section 5.6.5
on page 55). All pixels having a distance above RejectionThreshold are not assigned to any class.

The operator returns a tuple of regions in ClassRegions. This tuple contains one region for each class. The
sequence of the classes corresponds to the sequence used for the training regions that were added to the classifier
with add_samples_image_class_knn.
The returned image DistanceImage, contains the distance of each pixel of the input image Image to its nearest
neighbor.

6.1.7 Classification Based on Look-Up Tables

A significant speed-up for the image segmentation can be obtained by applying a classification that is based on
look-up tables (LUT). That is, you store the content of a trained classifier into a LUT und use the LUT-accelerated
classifier instead of the original classifier for the classification. The approach is as follows:
First, you create an MLP, SVM, GMM, or k-NN classifier, add samples to it, and apply the training as described
for the basic image segmentation in the previous sections.
Then, the LUT-accelerated classifier is created using the operator create_class_lut_mlp,
create_class_lut_svm, create_class_lut_gmm, or create_class_lut_knn, respectively. Here, you
insert the handle of the trained classifier and you can adjust the following parameters.

• ’bit_depth’ (for all classifiers)

The parameter ’bit_depth’ describes the number of bits used from the pixels. It controls the storage
requirement of the LUT-accelerated classifier and the runtime needed for the LUT-accelerated classification.
A byte image has a bit depth of 8. If the bit depth is set to a value of 7 or 6, the storage requirement can be
reduced. But note that a bit depth that is smaller than the bit depth of the image will usually lead to a lower
accuracy of the classification.
• ’class_selection’ (for MLP, SVM, and GMM classifiers)
the parameter ’class_selection’ is used to control the accuracy and the runtime needed for the creation
of the LUT-accelerated classifier. A higher accuracy slows down the runtime and a lower accuracy leads to
a speed-up. The value of ’class_selection’ is ignored if the bit depth of the LUT is maximal.
• ’rejection_threshold’ (for MLP, GMM, and k-NN classifiers)
The parameter ’rejection_threshold’ corresponds to the rejection threshold that is described for the
basic image segmentation in section [Link] for MLP, in section [Link] for GMM, and in section [Link] for
k-NN.

For the actual image segmentation, the operator classify_image_class_lut is applied instead of the
operator classify_image_class_mlp, classify_image_class_svm, classify_image_class_gmm, or
classify_image_class_knn, respectively. Note that the number of channels of the image that has to be clas-
sified must correspond to the value specified for the dimension of the feature space when creating the original
classifier, i.e., the value of NumInput in create_class_mlp, NumFeatures in create_class_svm, NumDim in
create_class_gmm, or NumDim in create_class_knn.
If you need the original trained classifier for further applications, you should store it to file (see section [Link] on
page 64). The LUT-accelerated classifier cannot be stored to file, as it needs a lot of memory and thus, it is more
suitable to create it again from the reused original classifier when starting a new application.
 6.1 Approach for MLP, SVM, GMM, and k-NN D-71

Figure 6.6: Fuses of different color are segmented using a LUT-accelerated GMM classifier.

The HDevelop example program %HALCONEXAMPLES%\hdevelop\Applications\Color-Inspection\

classify_fuses_gmm_based_lut.hdev uses a LUT-accelerated classifier that is based on a trained GMM
classifier to segment fuses of different color (see figure 6.6).
First, the ROIs that serve as training samples for different color classes are selected and concatenated into the tuple
Classes. Then, a GMM classifier is created, the training samples are added, and the classifier is trained. Until
now, the approach is the same as for the image segmentation that was described in the previous sections.

create_class_gmm (3, 5, 1, 'full', 'none', 3, 42, GMMHandle)

add_samples_image_class_gmm (Image, Classes, GMMHandle, 0)
train_class_gmm (GMMHandle, 100, 0.001, 'training', 0.001, Centers, Iter)

In contrast to the basic image segmentation, the training result, i.e., the classifier, is then stored in a LUT-
accelerated classifier using create_class_lut_gmm.

Segmentation
create_class_lut_gmm (GMMHandle, ['bit_depth', 'rejection_threshold'], [6, \
0.03], ClassLUTHandle)

Now, for each image that has to be classified, the operator classify_image_class_lut is applied to segment
the images based on the LUT of the new classifier.

for Img := 0 to 3 by 1
read_image (Image, ImageRootName + Img)
classify_image_class_lut (Image, ClassRegions, ClassLUTHandle)
endfor

The HDevelop example program %HALCONEXAMPLES%\hdevelop\Segmentation\Classification\

classify_image_class_lut.hdev compares the runtime needed by MLP, SVM, GMM, and k-NN clas-
sification using the basic image segmentation on the one hand and the LUT-accelerated classification on the other
hand. It exemplarily shows that the online part of the LUT-accelerated classification is significantly faster than
that of the basic image segmentation. But note that it needs also additional memory and the runtime for the
offline part is increasing. Because of that the LUT-accelerated classification is restricted to feature spaces with !
a maximum of three dimensions. Figure 6.7 shows the interdependencies between the number of classes, the bit
depth, and the required memory for a three-channel image. Furthermore, depending on the selected bit depth of
the look-up table, the accuracy of the classification may be decreased. Summarized, if a three-channel image
has to be segmented, a speed-up of the online part is required, and the runtime for the offline part is not critical,
LUT-accelerated classification is a good alternative to the basic image segmentation.
D-72 Classification for Image Segmentation

Number of classes Bit depth Needed memory

256 3 x 8 Bit 32 MB
1 3 x 8 Bit 2 MB
10 3 x 6 Bit 0,125 MB

Figure 6.7: Memory needed for a LUT-accelerated classifier for three-channel images with different numbers of
classes and different bit depths.

6.2 Approach for a Two-Channel Image Segmentation

For two-channel images a simple and very fast pixel classification can be applied using the operator
class_2dim_sup. As we already learned, a class is defined as a well-defined part of the feature space and the
dimension of the feature space depends on the number of features used for the classification. In case of a two-
channel image and a pure pixel based classification, the classification is based only on the two gray values that are
assigned to each pixel position and thus the two-dimensional feature space can be visualized in a 2D graph or in a
2D image, respectively. There, for each position of the image or a specific image region the gray value of the first
image ImageCol is used as column coordinate and the gray value of the second image ImageRow is used as row
coordinate (see figure 6.8, left). A class is defined by the feature space of a manually specified image region.
The general approach using class_2dim_sup for a two-channel image segmentation is as follows: you first
specify a class by a region in the two-channel image that is typical for the class. Then, you apply the oper-
ator histo_2dim, which needs the two channels and the specified image region as input and returns the two-
dimensional histogram, i.e., an image in which the position of a pixel is built by the combination of the gray values
of the two channels and its gray value is defined by the frequency of the specific gray value combination. To
extract the essential region of the feature space, a threshold is applied. Now, the result can be preprocessed for
generalization purposes (see figure 6.8, right), e.g., by a morphological operation like closing_circle. For the
actual classification, the feature space region and the two channels of the image that has to be classified are used
as input for the operator class_2dim_sup. The returned region RegionClass2Dim consists of all pixels in the
classified two-channel image, for which the gray value distribution is similar to the gray value distribution of the
training region, i.e., for which the position in the feature space lies inside the preprocessed feature space region
that defines the class.
g(ImageRow) g(ImageRow)

255 255

0 g(ImageCol) 0 g(ImageCol)

0 255 0 255

Figure 6.8: Positions of the 2D feature space for a supervised classification: (left) gray values of the two images in
a 2D graph, (right) feature space region (class) after generalization.

The example %HALCONEXAMPLES%\hdevelop\Segmentation\Classification\class_2dim_sup.hdev

shows the approach for the segmentation of a two-channel image that contains several capacitors. First the color
image is decomposed into its three channels, so that two of them can be accessed for the classification. Then,
a sample region for one of the capacitors is defined and used as input for histo_2dim, which together with
threshold and closing_circle is used to derive a generalized feature space region. Finally, the two channels
of the image and the trained feature space region are used by class_2dim_sup to get the region with all pixels
that correspond to the trained feature space.
 6.3 Approach for Euclidean and Hyperbox Classification D-73

read_image (Image, 'ic')

decompose3 (Image, Red, Green, Blue)
gen_rectangle1 (Pattern, 362, 276, 371, 298)
histo_2dim (Pattern, Red, Blue, Histo2Dim)
threshold (Histo2Dim, Features, 1, 255)
closing_circle (Features, FeaturesClosed, 11.5)
class_2dim_sup (Red, Blue, FeaturesClosed, RegionClass2Dim)

This image segmentation approach is very fast. To select the two channels of a multi-channel image that should be
used for the classification, different approaches exist. For color images, i.e., three-channel ’rgb’ images, you can,
e.g., apply a color transformation using trans_from_rgb to transform the ’rgb’ image into a, e.g., ’hsi’ image.
This image contains one channel for the ’hue’, one for the ’saturation’, and one for the ’intensity’ of the pixels.
When using the ’hue’ and ’saturation’ channels for the classification, you obtain a classifier that is invariant to
illumination changes (see also Solution Guide I, chapter 14 on page 147 for color transformations). For arbitrary
multi-channel images you can also transform your image by a principal component analysis using the operator
principal_comp. Then, the first two channels of the returned image are the channels with the largest information
content and thus are predestinated to be input to the two-channel image segmentation.

6.3 Approach for Euclidean and Hyperbox Classification

For the simple image segmentation of multi-channel images, class_ndim_norm is available. It can be applied
either using a hyperbox or an Euclidean metric (for hyperbox and Euclidean classification see section 3.2 on page
16). It can be applied following the general idea of hyperbox classification.
For learn_ndim_norm, an overlap between the classes ’Foreground’ and ’Background’ is allowed. This
has its effect on the return value Quality. The larger the overlap, the smaller the value. Note that the operator
class_ndim_norm is very efficient. Compared, e.g., to the classic image segmentation using GMM (section 6.1.5
on page 68), it is significantly faster (approximately by factor 3). Compared to a LUT-accelerated classification it
has the advantage that the storage requirements are low and the feature space can be easily visualized. Thus, if the
classes are build by compact clusters, it is a good alternative.

[Link] Classification with class_ndim_norm

When segmenting an image using class_ndim_norm, you do not explicitly create and destroy a classifier, but
immediately apply the training using learn_ndim_norm. Instead of storing the training results in a classifier and

Segmentation
using the classifier as input for the classification, the training returns explicit information about the centers and
radii of the clusters related to the trained patterns and this information is used as input for the classification, which
is applied with class_ndim_norm. You can choose between a classification using hyperboxes and a classification
using hyperspheres (Euclidean classification).
With learn_ndim_norm you generate the classification clusters from the region Foreground. The region
Background can be used to define a rejection class, but may also be empty (an empty region can be created
by gen_empty_region). Note that the rejection class does not influence the clustering, but can be used to detect
problems that might occur because of overlapping classes.
To choose between the two available clustering approaches, the parameter Metric is used. It can be either set to
’euclid’, which uses a minimum distance algorithm (n-dimensional hyperspheres) or to ’maximum’, which uses
n-dimensional hyperboxes to built the clusters. Metric must be set to the same value for the training as well as for
the classification. The Euclidean metric usually yields the better results but needs more run time.
The parameter Distance describes the minimum distance between two cluster centers and thus determines the
maximum value allowed for the output parameter Radius. Note that the cluster centers depend on the sequence
used to add the training samples (pixels). Thus, it is recommended to select a small value for Distance. Then,
the (small) hyperboxes or hyperspheres can approximate the feature space well. But simultaneously, the runtime
during classification increases.
The ratio of the number of pixels in a cluster to the total number of pixels (in percent) must be larger than the
value of the parameter MinNumberPercent, otherwise the cluster is not returned. MinNumberPercent serves to
eliminate outliers in the training set. If it is chosen too large many clusters are suppressed.
D-74 Classification for Image Segmentation

As result of the operator, the parameter Radius returns the minimum distance between two cluster centers, i.e.,
radii for hyperspheres or half edge lengths for hyperboxes, and Center returns the coordinates of the cluster
centers. Furthermore, the parameter Quality returns the quality of the clustering, i.e., a measure of overlap
between the rejection class and the classifier classes. Values larger than 0 denote the corresponding ratio of overlap.
If no rejection region is given, its value is set to 1. The regions in Background do not influence the clustering.
They are merely used to check the results that can be expected.
When classifying a multi-channel image with class_ndim_norm, you set the same metric (Metric) used also
for the training and set the parameter SingleMultiple to ’single’ if one region has to be generated or to
’multiple’ if multiple regions have to be generated for each cluster. Additionally, Radius and Center, which
were returned by learn_ndim_norm, are inserted. The result of class_ndim_norm is returned in Regions, which
either contains a single region or a tuple of regions, depending on the value of SingleMultiple.
The example %HALCONEXAMPLES%\hdevelop\Segmentation\Classification\class_ndim_norm.hdev
shows how to apply an image segmentation with class_ndim_norm. The image is read and within the image a
region for the class to be trained and an empty region for the rejection class are generated and used as input for the
training with learn_ndim_norm. The classification is then applied with class_ndim_norm. The result of the
image segmentation is shown in figure 6.9.

read_image (Image, 'ic')

gen_rectangle1 (Region, 360, 198, 369, 226)
gen_empty_region (EmptyRegion)
learn_ndim_norm (Region, EmptyRegion, Image, 'euclid', 10, 0.01, Radius, \
Center, Quality)
class_ndim_norm (Image, Regions, 'euclid', 'multiple', Radius, Center)

Figure 6.9: Result of image segmentation using class_ndim_norm (returned regions marked in white).
 Classification for Optical Character Recognition (OCR) D-75

OCR
Chapter 7

Classification for Optical Character

Recognition (OCR)

If classification is used for optical character recognition (OCR), individual regions are first extracted from the
image by a segmentation and then, with the help of some region features, assigned to classes that typically (but
not necessarily) represent individual characters or numbers. Approaches that are suitable for this feature-based
OCR comprise the MLP, the SVM, and the k-NN classifiers. A hyperbox classifier is also provided, but is not
recommended anymore. With Deep OCR a further OCR method is provided by HALCON (see Solution Guide I,
chapter 19 on page 229). This deep-learning-based method is not a classifier and as a consequence ignored in the
explanations within this manual.
In section 7.1 the general approach for OCR is illustrated by an example that trains and reads the characters ’A’
to ’G’. In section 7.2 the steps of OCR and the involved operators are listed for a brief overview. The parameters
used for the basic OCR operators are introduced in section 7.3 for MLP, section 7.4 for SVM, section 7.5 for k-NN,
and section 7.6 for CNNs. Section 7.7 finally lists all features that are available for OCR.
In general, when trying OCR for your application, it is recommended to first try the pretrained CNN-based OCR
font Universal and evaluate the result before you try any other OCR classification approaches, since the CNN-
based OCR classifier can generalize quite well.

7.1 General Approach

Figure 7.1 shows the general approach for OCR. Typically, the approach is divided into an offline and an online
process. The offline process comprises the training of the font, i.e., regions that represent characters or numbers
(in the following just called ’characters’) are extracted and stored together with the corresponding character names
in training files. The content of a training file optionally can be accessed again. The access is needed for different
reasons. First, they can be used to find errors that occurred during the training, which is needed on one hand for
your general quality assurance and on the other hand for the correspondence with your HALCON support team
(if needed), and second, you can reuse the contained information for the case that you want to apply a similar
application in the future.
Now, the training files are used to train the font. To access the font in the later online process, the classifier is
written into a font file.
If you want to read a font that is rather common and you want to use the MLP approach, you can also use one
of the pretrained fonts provided by HALCON (see Solution Guide I, chapter 13 on page 133 for illustrations of
the provided fonts). The pretrained fonts are stored in the subdirectory ocr of the directory where you installed
HALCON. Then, you can skip the offline training process. For SVM and k-NN classifiers, no pretrained fonts are
available.
In the online process, the font file is read so that the classifier can be accessed again. Then, the regions of unknown
characters are extracted, most suitably by the same method that was used within the offline training process, and
classified, i.e., the characters are read.
D-76 Classification for Optical Character Recognition (OCR)

Train New Classifier (Font) or Use Pretrained Font

Segment Regions of Known Characters

Store Regions and Character Names to Training Files
Offline
Check Correctness of Training Files
Train Classifier (Font)
Store Classifier to File

Read Characters
Online

Read Classifier (Font) from File

Segment Regions of Unknown Characters
Classify Regions (Read Characters)

Figure 7.1: The basic steps of OCR.

In the following, we illustrate the general approach for OCR classification with the examples %HALCONEXAMPLES%\
solution_guide\classification\train_characters_ocr.hdev and %HALCONEXAMPLES%\
solution_guide\classification\classify_characters_ocr.hdev. These examples show how the
characters ’A’, ’B’, ’C’, ’D’, ’E’, ’F’, and ’G’ are first trained (see figure 7.2) and then read (see figure 7.3)
with an SVM-based OCR classification. Note that the number of classes as well as the number of training samples
is very small as the example is used only to demonstrate the general approach. Typically, a larger number of
classes is trained with OCR and a lot of samples and probably a different set of features are needed to get a robust
classification.
The examples use SVM-based OCR classification. For MLP and k-NN, the general approach and the operators are
similar. Then, you mainly have to replace ’svm’ by ’mlp’ or ’knn’, respectively in the specific operator names
and adjust different parameters. The specific parameters are explained in more detail in section 7.3 on page 80 for
MLP, in section 7.4 on page 83 for SVM, and in section 7.5 on page 86 for k-NN.

Figure 7.2: Training images for the characters ’A’, ’B’, ’C’, ’D’, ’E’, ’F’, and ’G’.

The program %HALCONEXAMPLES%\solution_guide\classification\train_characters_ocr.hdev starts

with the creation of an SVM-based OCR classifier using the operator create_ocr_class_svm. Here, the most
important parameters are adjusted. The width and height of a normalized character is defined (the reason for this
 7.1 General Approach D-77

is explained in more detail in section 7.3.1 on page 80) and the mode for the interpolation that is used to scale
the characters to the average character size is set. Further, the features that should be calculated to get the feature
vector are selected. OCR can be applied for a broad set of features, which is listed in section 7.7 on page 89.

OCR
The default features are ’ratio’ and ’pixel_invar’. Here, the default does not lead to a satisfying result, so
we use the features ’convexity’, ’num_holes’, ’projection_horizontal’, and ’projection_vertical’
instead. Now, the names of the available classes (the characters A to G) are assigned, which were previously
stored in the tuple ClassNames. Furthermore, some SVM specific parameters are adjusted that are described in
more detail for the general classification in section 5.4 on page 42. The output of create_ocr_class_svm is the
handle of the classifier (OCRHandle), which is needed for the following classification steps.

ClassNames := ['A', 'B', 'C', 'D', 'E', 'F', 'G']

create_ocr_class_svm (8, 10, 'constant', ['convexity', 'num_holes', \
'projection_horizontal', 'projection_vertical'], \
ClassNames, 'rbf', 0.02, 0.05, 'one-versus-one', \
'normalization', 10, OCRHandle)

For the training of the characters, the training images are read and the regions of the characters are extracted via a
blob analysis within the procedure get_regions. Alternatively you can use also a combination of the operators
segment_characters and select_characters to extract regions.
In the program, the first region is added together with its class name (which is obtained from the tuple ClassName)
to a new training file with write_ocr_trainf. All following regions and their corresponding class names are
appended to this training file with append_ocr_trainf.

for I := 1 to 7 by 1
read_image (Image, 'ocr/chars_training_' + I$'.2d')
get_regions (Image, SortedRegions)
count_obj (SortedRegions, NumberObjects)
for J := 1 to NumberObjects by 1
select_obj (SortedRegions, ObjectSelected, J)
if (I == 1 and J == 1)
write_ocr_trainf (ObjectSelected, Image, ClassNames[J - 1], \
'train_characters_ocr.trf')
else
append_ocr_trainf (ObjectSelected, Image, ClassNames[J - 1], \
'train_characters_ocr.trf')
endif
endfor
endfor

After all samples were added to the training file, the operator read_ocr_trainf is applied to check if the training
samples and the corresponding class names were correctly assigned within the training file. In a short for-loop,
the individual characters and their corresponding class names are visualized. Note that the index for iconic objects
(Characters) starts with 1 and that of numeric objects (CharacterNames) with 0.

read_ocr_trainf (Characters, 'train_characters_ocr.trf', CharacterNames)

count_obj (Characters, NumberCharacters)
for I := 1 to NumberCharacters by 1
select_obj (Characters, CharacterSelected, I)
dev_display (CharacterSelected)
disp_message (WindowHandle, CharacterNames[I - 1], 'window', 10, 10, \
'black', 'true')
endfor

Then, the OCR classifier is trained with trainf_ocr_class_svm, which needs the training file as input. For
SVM, the number of support vectors obtained from the training can be reduced to enhance the speed of the later
classification. This is done with reduce_ocr_class_svm. The resulting handle is stored in a font file with
write_ocr_class_svm for later access. For MLP the handle obtained directly by the training would be stored.

trainf_ocr_class_svm (OCRHandle, 'train_characters_ocr.trf', 0.001, \

'default')
reduce_ocr_class_svm (OCRHandle, 'bottom_up', 2, 0.001, OCRHandleReduced)
write_ocr_class_svm (OCRHandleReduced, 'font_characters_ocr')
D-78 Classification for Optical Character Recognition (OCR)

Now, the example program %HALCONEXAMPLES%\solution_guide\classification\

classify_characters_ocr.hdev is used to read unknown characters of the same font type used for the
training.
A font is read from file with read_ocr_class_svm. Then, the images with unknown characters are read, and the
regions that most probably represent characters are extracted. If possible, the method to extract the regions should
be the same for the offline and online process. Of course this advice can only be followed if the methods used in
the offline process are known, i.e., it can not be followed when using pretrained fonts.
The extracted regions are then classified, i.e., the characters are read. In the example, we read all regions si-
multaneously with do_ocr_multi_class_svm. Alternatively, you can also read the regions individually with
do_ocr_single_class_svm. Then, not only the best class for each region is returned, but also the second best
(and third best etc.) class can be obtained, which might be suitable when having overlapping classing. But if only
the best class is of interest, do_ocr_multi_class_svm is faster and therefore recommended.

Figure 7.3: Classifying the characters ’A’, ’B’, ’C’, ’D’, ’E’, ’F’, and ’G’ with OCR.

read_ocr_class_svm ('font_characters_ocr', OCRHandle)

for I := 1 to 3 by 1
read_image (Image, 'ocr/chars_' + I$'.2d')
get_regions (Image, SortedRegions)
do_ocr_multi_class_svm (SortedRegions, Image, OCRHandle, Classes)
area_center (SortedRegions, AreaCenter, Row, Column)
count_obj (SortedRegions, NumberObjects)
disp_message (WindowHandle, Classes, 'window', Row - 100, Column, \
'black', 'true')
if (I < 3)
endif
endfor
endfor

7.2 Involved Operators (Overview)

This section gives a brief overview on the operators that are provided for OCR. In particular, first the operators for
the basic steps and then the advanced operators used for OCR are introduced.
 7.2 Involved Operators (Overview) D-79

[Link] Basic Operators

The basic steps apply the following operators in the following sequence.

OCR
1. Create an OCR classifier using create_ocr_class_mlp, create_ocr_class_svm, or
create_ocr_class_knn.
2. Extract the regions that represent the characters that have to be trained.
3. Store the samples, i.e., the training regions and the corresponding class names to a training file. This can be
done in different ways. Either
• store all samples at once using write_ocr_trainf. Then the regions as well as the corresponding
class names have to be available in a tuple. Or
• successively add the individual regions (characters) and their corresponding class names to the training
file using append_ocr_trainf.
• Write characters into a training file with write_ocr_trainf_image. That is, regions, represent-
ing characters, including their gray values (region and pixel) and the corresponding class name are
written into a file. An arbitrary number of regions within one image is supported. In contrast to
write_ocr_trainf one image per character is passed. The domain of this image defines the pixels
which belong to the character. The file format can be defined by the parameter ’ocr_trainf_version’ of
the operator set_system.
Additionally, several training files can be concatenated with concat_ocr_trainf.
4. Read the training characters from the training file and convert them into images with read_ocr_trainf to
check the correctness of the training file content.
5. Train the OCR classifier with trainf_ocr_class_mlp, trainf_ocr_class_svm, or
trainf_ocr_class_knn.
6. Write the OCR classifier to a font file with write_ocr_class_mlp (default file extension: .omc),
write_ocr_class_svm (default file extension: .osc), or write_ocr_class_knn (default file extension:
.onc).
7. Read the OCR classifier from the font file with read_ocr_class_mlp, read_ocr_class_svm,
read_ocr_class_knn, or read_ocr_class_cnn.
8. Extract the regions of the characters that have to be classified according to the trained font.
9. Classify the regions of the characters to be classified. Here, you have different possibilities:
• Classify multiple characters with an OCR classifier with do_ocr_multi_class_mlp,
do_ocr_multi_class_svm, do_ocr_multi_class_knn, or do_ocr_multi_class_cnn.
• Classify a single character with an OCR classifier with do_ocr_single_class_mlp,
do_ocr_single_class_svm, do_ocr_single_class_knn, or do_ocr_single_class_cnn.

[Link] Advanced Operators

Besides the basic steps of an OCR classification, some additional steps and operators can be applied if suitable. In
particular, you can

• compute the features of a character with get_features_ocr_class_mlp,

get_features_ocr_class_svm, or get_features_ocr_class_knn,
• return the parameters of an OCR classifier with get_params_ocr_class_mlp,
get_params_ocr_class_svm, get_params_ocr_class_knn, or get_params_ocr_class_cnn,
• compute the information content of the preprocessed feature vectors of an OCR classifier with
get_prep_info_ocr_class_mlp or get_prep_info_ocr_class_svm (if Preprocessing was set to
’principal_components’ or ’canonical_variates’; note that this kind of operator is not available
for k-NN classifiers, because they do not provide the respective preprocessing options),
D-80 Classification for Optical Character Recognition (OCR)

• query which characters are stored in a training file with read_ocr_trainf_names,

• read training specific characters from files and convert them to images with read_ocr_trainf_select, or
• classify a related group of characters with an OCR classifier with do_ocr_word_mlp,
do_ocr_word_svm, do_ocr_word_cnm, do_ocr_word_knn, or do_ocr_word_cnn. This is an al-
ternative to do_ocr_multi_class_mlp, do_ocr_multi_class_svm, do_ocr_multi_class_cnn, or
do_ocr_multi_class_knn and is suitable when searching for specific words or regular expressions that
are specified in a lexicon that has been created with create_lexicon or imported with import_lexicon.

Besides these operators, some operators are provided that are available only for SVM:

• To enhance the speed, the trained SVM-based OCR classifier can be approximated with a reduced number
of support vectors by applying reduce_ocr_class_svm after the training.
• The number of support vectors that are stored within the SVM-based OCR classifier is returned with
get_support_vector_num_ocr_class_svm.
• The index of a support vector from a trained SVM-based OCR classifier is returned with
get_support_vector_ocr_class_svm.

In the following, the parameters for the basic operators are introduced and tips for their adjustment are provided.

7.3 Parameter Setting for MLP

The following sections introduce you to the parameters that have to be set for the basic operators needed for an
MLP-based classification for OCR.

7.3.1 Adjusting create_ocr_class_mlp

An MLP classifier for OCR is created using create_ocr_class_mlp. Here, most of the important parameters
have to be set.

Parameter WidthCharacter / HeightCharacter

Like for the general classification described in section 5 on page 31, OCR uses a set of features to classify regions
into classes, which in this case correspond to specific characters and numbers. Some of the features that can be
used for OCR are gray value features for which the number of returned features varies dependent on the region’s
size. As a classifier requires a constant number of features, i.e., the dimension of the feature vector has to be the
same for all training samples and all regions to be classified, the region of a character has to be transformed (scaled)
to a standard size. This size is determined by WidthCharacter and HeightCharacter. In most applications,
sizes between 6x8 and 10x14 should be used. As a rule of thumb, the values may be small if only few characters
have to be distinguished but large when classifying characters with complex shapes (e.g., Japanese signs).

Parameter Interpolation
The input parameter Interpolation is needed to control the transformation of the region to the size specified
with WidthCharacter and HeightCharacter. Generally, when transforming a region, transformed points will
lie between discrete pixel coordinates. To assign each point to its final pixel coordinate and additionally to avoid
aliasing, which typically occurs for scaled regions, an appropriate interpolation scheme is needed. Three types
of interpolation with different quality and speed properties are provided by HALCON and can be selected for the
parameter Interpolation:

• ’nearest_neighbor’
If ’nearest_neighbor’ is selected, a nearest-neighbor interpolation is applied. There, the gray value is
determined from the nearest pixel’s gray value. This interpolation scheme is very fast but may lead to a low
interpolation quality.
 7.3 Parameter Setting for MLP D-81

• ’bilinear’
If ’bilinear’ is selected, a bilinear interpolation is applied. There, the gray value is determined from the
four nearest pixels through bilinear interpolation. This interpolation scheme is of medium speed and quality.

OCR
Do not use it if the characters in the image appear larger than WidthCharacter times HeightCharacter.
In this case, the interpolation method ’constant’ or ’weighted’ should be used.
• ’constant’
If ’constant’ is selected, a bilinear interpolation is applied. There, the gray value is determined from
the four nearest pixels through bilinear interpolation. If the transformation contains a scaling with a scale
factor smaller than 1, a kind of mean filter is used to prevent aliasing effects. This interpolation scheme is of
medium speed and quality.
• ’weighted’
If ’weighted’ is selected, again a bilinear interpolation is applied, but now, aliasing effects for a scaling
with a scale factor smaller than 1 are prevented by a kind of Gaussian filter instead of a mean filter. This
interpolation scheme is slow but leads to the best quality.

The interpolation should be chosen such that no aliasing effects occur in the transformation. For most applications,
Interpolation should be set to ’constant’.

Parameter Features
The input parameter Features specifies the features that are used for the classification. Features can contain a
tuple of several feature names. Each of these feature names results in one or more features to be calculated for the
classifier. That is, the length of the tuple in Features is similar or smaller than the dimension of the final feature
vector. In section 7.7 on page 89 all features that are available for OCR are listed. To classify characters, in most
cases the ’default’ setting can be used. Then, the features ’ratio’ and ’pixel_invar’ are selected. Note
that the selection of the features significantly influences the quality of the classification.

Parameter Characters
The input parameter Characters contains a tuple with the names of the characters that will be trained. Each name
must be passed as a string. The number of elements in the tuple determines the number of available classes.

Parameter NumHidden
The input parameter NumHidden follows the same rules as provided for the corresponding operator used to create
an MLP classifier for a general classification (see section 5.3.1 on page 37).

Parameters Preprocessing / NumComponents

The parameters Preprocessing and NumComponents follow the same rules as provided for the corresponding
operator used to create an MLP classifier for a general classification (see section 5.3.1 on page 37). The only excep-
tion is that for the OCR classification the features are already approximately normalized. Thus, Preprocessing
can typically be set to ’none’.
If Preprocessing is set to ’principal_components’ or ’canonical_variates’ you can use the operator
get_prep_info_ocr_class_mlp to determine the optimum number of components as described for the general
classification in section 5.3.1 on page 38.

Parameter RandSeed
The parameter RandSeed follows the same rules as provided for the corresponding operator used to create an MLP
classifier for a general classification (see section 5.3.1 on page 37).

Parameter OCRHandle
The output parameter OCRHandle is the handle of the classifier that is needed and modified throughout the follow-
ing classification steps.
D-82 Classification for Optical Character Recognition (OCR)

7.3.2 Adjusting write_ocr_trainf / append_ocr_trainf

After creating a classifier and segmenting the regions for the characters of known classes, i.e., character names, the
samples must be stored to a training file. Here, the same operators are used for MLP and SVM classifiers.
Different operators are provided for storing training samples to file. You can either store all samples to file in
one step by inserting a tuple containing all regions and a tuple containing all corresponding class names to the
operator write_ocr_trainf. Or you can successively append single training samples to the training file using
append_ocr_trainf. If you choose the latter, be aware that the training file is extended every time you run
the program, i.e., it is not created anew. So, if you want to successively add samples, it is recommended to use
write_ocr_trainf for the first sample and append_ocr_trainf for all following samples. The operators are
applied as follows:

• Store training samples to a new training file:

When storing all training samples simultaneously into a file using write_ocr_trainf, you have to assign
a tuple of regions that represent characters to the parameter Character. The image that contains the regions
must be set in Image so that knowledge about the gray values within the regions is available. In Class a
tuple of class names that correspond to the regions with the same tuple index must be inserted. Finally, you
specify the name and path of the stored training file in FileName.
• Append training samples to a training file:
When successively storing individual training samples into a file using append_ocr_trainf, the same
parameters as for write_ocr_trainf have to be set. But in contrast to the operator write_ocr_trainf
the characters are appended to an existing file using the same training file format. If the file does not exist, a
new file is generated.

If no file extension is specified in FileName the extension ’.trf’ is appended to the file name. The version of
the file format used for writing data can be defined by the parameter ’ocr_trainf_version’ of the operator
set_system.
If you have several training files that you want to combine, you can concatenate them with the operator
concat_ocr_trainf.

7.3.3 Adjusting trainf_ocr_class_mlp

trainf_ocr_class_mlp trains the OCR classifier OCRHandle with the training characters stored in the
OCR training file given by TrainingFile. The remaining parameters MaxIterations, WeightTolerance,
ErrorTolerance, Error, and ErrorLog have the same meaning as introduced for the training of an MLP classi-
fier for a general classification (see section 5.3.3 on page 40).

7.3.4 Adjusting do_ocr_multi_class_mlp

With do_ocr_multi_class_mlp multiple characters can be classified in a single call. Typically, this is faster
than successively applying do_ocr_single_class_mlp, which classifies single characters, in a loop. However,
do_ocr_multi_class_mlp can only return the best class of each character. If the second best class is needed,
e.g., because the classes significantly overlap (see section 5.3.5 on page 42 for the possible outliers related to the
confidence values of MLP classifications), do_ocr_single_class_cnn should be used instead. The following
parameters have to be set for do_ocr_multi_class_mlp:

Parameter Character
The input parameter Character contains a tuple of regions that have to be classified.

Parameter Image
The input parameter Image contains the image that provides the gray value information for the regions that have
to be classified.
 7.4 Parameter Setting for SVM D-83

Parameter OCRHandle
The input parameter OCRHandle is the handle of the classifier that was trained with trainf_ocr_class_mlp.

OCR
Parameter Class
The output parameter Class returns the result of the classification, i.e., a tuple of character names that correspond
to the input regions that were given in the tuple Character. If the result is ’\0x1A’ or ’\032’, the respective region
has been classified as rejection class.

Parameter Confidence
The output parameter Confidence returns the confidence value for the classification. Note, that for the confidence
of MLP classifications, outliers are possible as described for the general classification in section 5.3.5 on page 42.

7.3.5 Adjusting do_ocr_single_class_mlp

Instead of using do_ocr_multi_class_mlp to add all samples in a single call, do_ocr_single_class_mlp can
be used to successively add samples. Then, besides the best class for a region, also the second best (and third best
etc.) class can be obtained. This may be suitable, if the class membership of a region is uncertain because of, e.g.,
overlapping classes. If only the best class for each region is searched for, do_ocr_multi_class_mlp is faster
and therefore recommended.

Parameter Character
The input parameter Character contains a single region that has to be classified.

Parameter Image
The input parameter Image contains the image that provides the gray value information for the region that has to
be classified.

Parameter OCRHandle
The input parameter OCRHandle is the handle of the classifier that was trained with trainf_ocr_class_mlp.

Parameter Num
The input parameter Num specifies the number of best classes to be searched for. Generally, Num is set to 1 if only
the class with the best probability is searched for, and to 2 if the second best class is also of interest, e.g., because
the classes overlap.

Parameter Class
The output parameter Class returns the result of the classification, i.e., the Num best character names that corre-
spond to the input region that was specified in Character. If the result is ’\0x1A’ or ’\032’, the respective region
has been classified as rejection class.

Parameter Confidence
The output parameter Confidence returns the Num best confidence values for the classification. Note, that for the
confidence of MLP classifications, outliers are possible as described for the general classification in section 5.3.5
on page 42.

7.4 Parameter Setting for SVM

The following sections introduce you to the parameters that have to be set for the basic operators needed for an
SVM-based classification for OCR.
D-84 Classification for Optical Character Recognition (OCR)

7.4.1 Adjusting create_ocr_class_svm

An SVM classifier for OCR is created using create_ocr_class_svm. Here, most of the important parameters
have to be set.

Parameters WidthCharacter / HeightCharacter

Like for the general classification described in section 5 on page 31, OCR uses a set of features to classify regions
into classes, which in this case correspond to specific characters and numbers. Some of the features that can be
used for OCR are gray value features for which the number of returned features varies dependent on the region’s
size. As a classifier requires a constant number of features, i.e., the dimension of the feature vector has to be the
same for all training samples and all regions to be classified, the region of a character has to be transformed (scaled)
to a standard size. This size is determined by WidthCharacter and HeightCharacter. In most applications,
sizes between 6x8 and 10x14 should be used.

Parameter Interpolation
The input parameter Interpolation is needed to control the transformation of the region to the size speci-
fied with WidthCharacter and HeightCharacter. Generally, when transforming a region, transformed points
will lie between discrete pixel coordinates. To assign each point to its final pixel coordinate and additionally to
avoid aliasing, which typically occurs for scaled regions, an appropriate interpolation scheme is needed. The four
types of interpolation that are provided by HALCON are ’nearest_neighbor’, ’bilinear’, ’constant’, and
’weighted’. The properties of the individual interpolation schemes were already introduced for the creation of
an MLP-based OCR classifier in section 7.3.1 on page 80. For most applications, Interpolation should be set
to ’constant’.

Parameter Features
The input parameter Features specifies the features that are used for the classification. Features can contain a
tuple of several feature names. Each of these feature names results in one or more features to be calculated for the
classifier. That is, the length of the tuple in Features is similar or smaller than the dimension of the final feature
vector. In section 7.7 on page 89 all features that are available for OCR are listed. To classify characters, in most
cases the ’default’ setting can be used. Then, the features ’ratio’ and ’pixel_invar’ are selected. Note
that the selection of the features significantly influences the quality of the classification.

Parameter Characters
The input parameter Characters contains a tuple with the names of the characters that will be trained. Each name
must be passed as a string. The number of elements in the tuple determines the number of available classes.

Parameters KernelType, KernelParam

The input parameters KernelType and KernelParam follow the same rules as provided for the corresponding
operator used to create an SVM classifier for a general classification (see section 5.4.1 on page 42).

Parameter Nu
The input parameter Nu follows the same rules as provided for the corresponding operator used to create an SVM
classifier for a general classification (see section 5.4.1 on page 42).

Parameter Mode
The input parameter Mode follows the same rules as provided for the corresponding operator used to create an
SVM classifier for a general classification (see section 5.4.1 on page 42).

Parameters Preprocessing / NumComponents

The parameters Preprocessing and NumComponents follow the same rules as provided for the corresponding
operator used to create a classifier for a general classification (see section 5.3.1 on page 37). For the sake of
numerical stability, Preprocessing can typically be set to ’normalization’. In order to speed up classification
 7.4 Parameter Setting for SVM D-85

time, ’principal_components’ or ’canonical_variates’ can be used, as the number of input features can
be significantly reduced without deterioration of the recognition rate.
If Preprocessing is set to ’principal_components’ or ’canonical_variates’ you can use the operator

OCR
get_prep_info_ocr_class_svm to determine the optimum number of components as described for the general
classification in section 5.3.1 on page 38.

Parameter OCRHandle
The output parameter OCRHandle is the handle of the classifier that is needed and modified throughout the follow-
ing classification steps.

7.4.2 Adjusting write_ocr_trainf / append_ocr_trainf

The approach to store the training samples into a training file is the same for MLP-based and SVM-based OCR
classification. See section 7.3.2 on page 82 for details.

7.4.3 Adjusting trainf_ocr_class_svm

trainf_ocr_class_svm trains the OCR classifier OCRHandle with the training characters stored in the OCR
training file given by TrainingFile. The remaining parameters Epsilon and TrainMode have the same meaning
as introduced for the training of an SVM classifier for a general classification (see section 5.4.3 on page 45).

7.4.4 Adjusting do_ocr_multi_class_svm

With do_ocr_multi_class_svm multiple characters can be classified in a single call. Typically, this is faster
than successively applying do_ocr_single_class_svm, which classifies single characters, in a loop. How-
ever, do_ocr_multi_class_svm can only return the best class of each character. If the second best class is
of interest, do_ocr_single_class_svm should be used instead. The following parameters have to be set for
do_ocr_multi_class_svm:

Parameter Character
The input parameter Character contains a tuple of regions that have to be classified.

Parameter Image
The input parameter Image contains the image that provides the gray value information for the regions that have
to be classified.

Parameter OCRHandle
The input parameter OCRHandle is the handle of the classifier that was trained with trainf_ocr_class_svm.

Parameter Class
The output parameter Class returns the result of the classification, i.e., a tuple of character names that correspond
to the input regions that were given in the tuple Character.

7.4.5 Adjusting do_ocr_single_class_svm

Instead of using do_ocr_multi_class_svm to add all samples in a single call, do_ocr_single_class_svm can
be used to successively add samples. Then, besides the best class for a region, also the second best (and third best
etc.) class can be obtained. If only the best class for each region is searched for, do_ocr_multi_class_mlp is
faster and therefore recommended.
D-86 Classification for Optical Character Recognition (OCR)

Parameter Character
The input parameter Character contains a single region that has to be classified.

Parameter Image
The input parameter Image contains the image that provides the gray value information for the region that has to
be classified.

Parameter OCRHandle
The input parameter OCRHandle is the handle of the classifier that was trained with trainf_ocr_class_svm.

Parameter Num
The input parameter Num specifies the number of best classes to be searched for. Generally, Num is set to 1 if only
the class with the best probability is searched for, and to 2 if the second best class is also of interest.

Parameter Class
The output parameter Class returns the result of the classification, i.e., the Num best character names that corre-
spond to the input region that was specified in Character.

7.5 Parameter Setting for k-NN

The following sections introduce you to the parameters that have to be set for the basic operators needed for an
k-NN-based classification for OCR.

7.5.1 Adjusting create_ocr_class_knn

A k-NN classifier for OCR is created using create_ocr_class_knn.

Parameters WidthCharacter / HeightCharacter

Like for the general classification described in section 5 on page 31, OCR uses a set of features to classify regions
into classes, which in this case correspond to specific characters and numbers. Some of the features that can be used
for OCR are gray value features for which the number of returned features varies dependent on the region’s size.
As a classifier requires a constant number of features, i.e., the dimension of the feature vector has to be constant
for all training samples and all regions to be classified, the region of a character has to be transformed (scaled) to
a standard size. This size is determined by WidthCharacter and HeightCharacter. In most applications, sizes
between 6x8 and 10x14 should be used.

Parameter Interpolation
The input parameter Interpolation is needed to control the transformation of the region to the size speci-
fied with WidthCharacter and HeightCharacter. Generally, when transforming a region, transformed points
will lie between discrete pixel coordinates. To assign each point to its final pixel coordinate and additionally to
avoid aliasing, which typically occurs for scaled regions, an appropriate interpolation scheme is needed. The four
types of interpolation that are provided by HALCON are ’nearest_neighbor’, ’bilinear’, ’constant’, and
’weighted’. The properties of the individual interpolation schemes were already introduced for the creation of
an MLP-based OCR classifier in section 7.3.1 on page 80. For most applications, Interpolation should be set
to ’constant’.

Parameter Features
The input parameter Features specifies the features that are used for the classification. Features can contain a
tuple of several feature names. Each of these feature names results in one or more features to be calculated for the
classifier. That is, the length of the tuple in Features is identical or smaller than the dimension of the final feature
vector. In section 7.7 on page 89 all features that are available for OCR are listed. To classify characters, in most
cases the ’default’ setting can be used. Then, the features ’ratio’ and ’pixel_invar’ are selected. Note
that the selection of the features significantly influences the quality of the classification.
 7.5 Parameter Setting for k-NN D-87

Parameter Characters
The input parameter Characters contains a tuple with the names of the characters that will be trained. Each name
must be passed as a string. The number of elements in the tuple determines the number of classes.

OCR
Generic parameters
The pair of input parameters GenParamName and GenParamValues is provided for future use only.

Parameter OCRHandle
The output parameter OCRHandle is the handle of the classifier that is needed and modified throughout the follow-
ing classification steps.

7.5.2 Adjusting write_ocr_trainf / append_ocr_trainf

The approach to store the training samples into a training file is the same for MLP-based and k-NN-based OCR
classification. See section 7.3.2 on page 82 for details.

7.5.3 Adjusting trainf_ocr_class_knn

trainf_ocr_class_knn trains the OCR classifier OCRHandle with the training characters stored in the OCR
training file given by TrainingFile. The generic parameters ’num_trees’ and ’normalization’ can be set
with GenParamName and GenParamValues. They have the same meaning as introduced for the training of an
k-NN classifier for a general classification (see section 5.6.3 on page 53).

7.5.4 Adjusting do_ocr_multi_class_knn

With do_ocr_multi_class_knn multiple characters can be classified in a single call. Typically, this is faster
than successively applying do_ocr_single_class_knn, which classifies single characters, in a loop. However,
do_ocr_multi_class_knn can only return the best class of each character. If the second best class is of inter-
est, as well, do_ocr_single_class_knn should be used instead. The following parameters have to be set for
do_ocr_multi_class_knn:

Parameter Character
The input parameter Character contains a tuple of regions that have to be classified.

Parameter Image
The input parameter Image contains the image that provides the gray value information for the regions that have
to be classified.

Parameter OCRHandle
The input parameter OCRHandle is the handle of the classifier that was trained with trainf_ocr_class_knn.

Parameter Class
The output parameter Class returns the result of the classification, i.e., a tuple of character names that correspond
to the input regions that were given in the tuple Character.

Parameter Confidence
The output parameter Confidence returns the confidence value for the classification.
D-88 Classification for Optical Character Recognition (OCR)

7.5.5 Adjusting do_ocr_single_class_knn

Instead of using do_ocr_multi_class_knn to classify all samples in a single call, do_ocr_single_class_svm

can be used to successively classify samples. Then, besides the best class for a given region, also the sec-
ond best (and third best etc.) class can be obtained. If only the best class for each region is searched for,
do_ocr_multi_class_knn will be faster and is therefore recommended.

Parameter Character
The input parameter Character contains a single region that has to be classified.

Parameter Image
The input parameter Image contains the image that provides the gray value information for the region that has to
be classified.

Parameter OCRHandle
The input parameter OCRHandle is the handle of the classifier that has been trained with trainf_ocr_class_knn.

Parameter NumClasses
The input parameter NumClasses specifies the maximum number of best classes to be returned. For example,
NumClasses is set to 1 if only the class with the best probability is searched for, and to 2 if the second best class is
also of interest. Note that less than NumClasses are returned, if the nearest NumNeighbors do not contain enough
different classes.

Parameter NumNeighbors
The parameter NumNeighbors defines the number of nearest neighbors that are determined during the classifica-
tion. The selection of a suitable value for NumNeighbors depends heavily on the particular application. Generally,
larger values of NumNeighbors lead to higher robustness against noise, smooth the boundaries between the classes,
and lead to longer runtimes during the classification.
In practice, the best way of finding a suitable value for NumNeighbors is indeed to try different values and to
select the value for NumNeighbors that yields the best classification results under the constraint of an acceptable
runtime.

Parameter Class
The output parameter Class returns the result of the classification, i.e., the maximally NumClasses best character
names that correspond to the input region that was specified in Character.

7.6 Parameter Setting for CNNs

Since training for CNN-based OCR classification is not available in HALCON, we only take a look at the operators
that are used to classify text using the pretrained font Universal (see Solution Guide I, section 18.8 on page 222).

7.6.1 Adjusting do_ocr_multi_class_cnn

With do_ocr_multi_class_cnn, multiple characters can be classified in a single call. Typically, this is faster
than successively applying do_ocr_single_class_cnn, which classifies single characters, in a loop. How-
ever, do_ocr_multi_class_cnn can only return the best class of each character. If the second best class is
of interest, do_ocr_single_class_mlp should be used instead. The following parameters have to be set for
do_ocr_multi_class_cnn:

Parameter Character
The input parameter Character contains a tuple of regions that have to be classified.
 7.7 OCR Features D-89

Parameter Image
The input parameter Image contains the image that provides the gray value information for the regions that have
to be classified.

OCR
Parameter OCRHandle
The input parameter OCRHandle is the handle of the classifier that was read with read_ocr_class_cnn.

Parameter Class
The output parameter Class returns the result of the classification, i.e., a tuple of character names that correspond
to the input regions that were given in the tuple Character.

7.6.2 Adjusting do_ocr_single_class_cnn

Instead of using do_ocr_multi_class_cnn to add all samples in a single call, do_ocr_single_class_cnn can
be used to successively add samples. Then, besides the best class for a region, also the second best (and third best
etc.) class can be obtained. If only the best class for each region is searched for, do_ocr_multi_class_cnn is
faster and therefore recommended.

Parameter Character
The input parameter Character contains a single region that has to be classified.

Parameter Image
The input parameter Image contains the image that provides the gray value information for the region that has to
be classified.

Parameter OCRHandle
The input parameter OCRHandle is the handle of the classifier that was read with read_ocr_class_cnn.

Parameter Num
The input parameter Num specifies the number of best classes to be searched for. Generally, Num is set to 1 if only
the class with the best probability is searched for, and to 2 if the second best class is also of interest.

Parameter Class
The output parameter Class returns the result of the classification, i.e., the Num best character names that corre-
spond to the input region that was specified in Character.

7.7 OCR Features

The features that determine the feature vector for an OCR specific classification are selected via the parameter
Feature in the operator create_ocr_class_mlp or create_ocr_class_svm, respectively. Note that some of
the features lead to more than one feature value, i.e., the dimension of the feature vector can be larger than the
number of selected feature types. The following feature types can be set individually or in combinations:

Feature ’anisometry’
Anisometry of the character. If Ra and Rb are the two radii of an ellipse that has the “same orientation” and the
“same side relation” as the input region, the anisometry is defined as:
Ra
anisometry = Rb
D-90 Classification for Optical Character Recognition (OCR)

Feature ’chord_histo’
Frequency of the runs per row. The number of returned features depends on the height of the pattern. Note that
this feature is not scale-invariant.

Feature ’compactness’
Compactness of the character. If L is the length of the contour and F the area of the region, the compactness is
defined as:
L2
OCR Feature compactness = 4πF

The compactness of a circle is 1. If the region is long or has holes, the compactness is larger than 1. The
compactness responds to the run of the contour (roughness) and to holes.

Feature ’convexity’
Convexity of the character. If Fc is the area of the convex hull and Fo the original area of the region, the convexity
is defined as:
Fo
convexity = Fc

The convexity is 1 if the region is convex (e.g., rectangle, circle etc.). If there are indentations or holes, the
convexity is smaller than 1.

Feature ’cooc’
Values of the binary co-occurrence matrices. A binary co-occurrence matrix describes how often the values 0
(outside the region) and 1 (inside the region) are located next to each other in a certain direction (0, 45, 90, 135
degrees). This numbers are stored in the co-occurrence matrix at the locations (0,0), (0,1), (1,0), and (1,1). Due
to the symmetric nature of the co-occurrence matrix, each matrix contains two independent entries, e.g., (0,0) and
(0,1). These two entries are taken from each of the four matrices. The feature type ’cooc’ returns eight features.

Feature ’foreground’
Fraction of pixels in the foreground.

Feature ’foreground_grid_16’
Fraction of pixels in the foreground in a 4x4 grid within the smallest enclosing rectangle of the character. The
feature type ’foreground_grid_16’ returns 16 features.

Feature ’foreground_grid_9’
Fraction of pixels in the foreground in a 3x3 grid within the smallest enclosing rectangle of the character. The
feature type ’foreground_grid_9’ returns nine features.

Feature ’gradient_8dir’
Gradients are computed on the character image. The gradient directions are discretized into 8 directions. The
amplitude image is decomposed into 8 channels according to these discretized directions. 25 samples on a 5x5
grid are extracted from each channel. These samples are used as features (200 features).

Feature ’height’
Height of the character before scaling the character to the standard size (not scale-invariant).

Feature ’moments_central’
Normalized central moments of the character. The feature type ’moments_central’ is invariant under other
affine transformations, e.g., rotation or stretching, and returns the four features psi1, psi2, psi3, and psi4.
 7.7 OCR Features D-91

Feature ’moments_gray_plane’
Normalized gray value moments and the angle of the gray value plane. This incorporates the gray value cen-
ter of gravity (g r;g c) together with the parameters α and β, which describe the orientation of the plane which

OCR
approximates the gray values. The feature type ’moments_gray_plane’ returns four features.

Feature ’moments_region_2nd_invar’
Normalized 2nd moments of the character. The feature type ’moments_region_2nd_invar’ returns the three
features µ11 , µ20 , and µ02 .

Feature ’moments_region_2nd_rel_invar’
Normalized 2nd relative moments of the character. The feature type ’moments_region_2nd_rel_invar’ returns
the two features phi1 and phi2.

Feature ’moments_region_3rd_invar’
Normalized 3rd moments of the character. The feature type ’moments_region_3rd_invar’ returns the four
features µ21 , µ12 , µ03 , and µ30 .

Feature ’num_connect’
Number of connected components.

Feature ’num_holes’
Number of holes.

Feature ’num_runs’
Number of runs in the region normalized by the area.

Feature ’phi’
Sine and cosine of the orientation of an ellipse that has the “same orientation” and the “same side relation” as the
input region. The feature type ’phi’ returns two features.

Feature ’pixel’
Gray values of the character. The number of returned features depends on the height and width of the pattern.

Feature ’pixel_binary’
Region of the character as a binary image. The number of returned features depends on the height and width of the
pattern.

Feature ’pixel_invar’
Gray values of the character with maximum scaling of the gray values. The number of returned features depends
on the height and width of the pattern.

Feature ’projection_horizontal’
Horizontal projection of the gray values, i.e., the mean values in the horizontal direction of the gray values of the
input image. The number of returned features depends on the height of the pattern.

Feature ’projection_horizontal_invar’
Maximally scaled horizontal projection of the gray values. The number of returned features depends on the height
of the pattern.
D-92 Classification for Optical Character Recognition (OCR)

Feature ’projection_vertical’
Vertical projection of the gray values, i.e., the mean values in the vertical direction of the gray values of the input
image. The number of returned features depends on the width of the pattern.

Feature ’projection_vertical_invar’
Maximally scaled vertical projection of the gray values. The number of returned features depends on the width of
the pattern.

Feature ’ratio’
Aspect ratio of the character.

Feature ’width’
Width of the character before scaling the character to the standard size (not scale-invariant).

Feature ’zoom_factor’
Difference in size between the character and the values of PatternWidth and PatternHeight (not scale-invariant).
Further information about the individual features and their calculation can be accessed via the Reference Manual
entries for create_ocr_class_mlp or create_ocr_class_svm, respectively.
 General Tips D-93

Chapter 8

Tips
General Tips

This sections provides you with some additional tips that may help you to optimize your classification application.
In particular, a method for optimizing the most critical parameters is introduced in section 8.1, the classification
of general region features with the OCR specific classification operators is described in section 8.2, and means to
visualize the feature space for low dimensional feature vectors (2D and 3D) are given in section 8.3.

8.1 Optimize Critical Parameters with a Test Application

To optimize the most critical parameters for a classification, different parameter values should be tested with the
available training data. To optimize the generalization ability of a classifier, the parameter optimization should be
combined with a cross validation. There, the training data is divided into typically five sub sets and the training is
performed rotative with four of the five sub sets and tested with the fifth sub set (see figure 8.1). Take care, that
the training data is uniformly distributed in the sub sets. That is, if for one class only five samples are available,
each sub set should contain one of it, and if for another class hundred samples are available, each sub set should
contain twenty of them. Note that a cross validation needs a lot of time as it is reasonable mainly for a very large
set of training data, i.e., for applications that are challenging because of the many variations inside the classes and
the overlaps between the classes.

Training 1 Set 1 Set 2 Set 3 Set 4 Set 5

Training 2 Set 1 Set 2 Set 3 Set 4 Set 5

Training 3 Set 1 Set 2 Set 3 Set 4 Set 5

Training 4 Set 1 Set 2 Set 3 Set 4 Set 5

Training 5 Set 1 Set 2 Set 3 Set 4 Set 5

Figure 8.1: Cross validation: The training data is divided into 5 sub sets. Each set is used once as test data (black)
that is classified by a classifier trained by the training data of the other 4 sub sets (gray).

The actual test application can be applied as follows:

• You first split up the training data into five uniformly distributed data sets.
• Then, you create a loop over the different parameter values that are to be tested, e.g., over different values
for NumHidden in case of an MLP classification. When adjusting two parameters simultaneously, e.g., the
Nu-KernelParam pair for SVM, you have to nest two loops into each other.
D-94 General Tips

• Within the (inner) loop, the cross validation is applied, i.e., each sub set of the training data is once classified
with a classifier that is trained by the other four sub sets using the tested parameters. The sum of the correctly
classified samples of the test dataset is stored so that later the sum of correct classifications can be assigned
to the corresponding tested parameter values.
• After testing all parameter values, you select the best result, i.e., the parameter values that led to the largest
number of correctly classified test samples within the test application are used for the actual classification
application.

For the cross validation, a number of five sub sets is sufficient. When increasing this number, no advantage
is obtained, but the training is slowed down significantly. The parameters for which such a test application is
reasonable mainly comprise NumCenters for GMM (then, the parameter CovarType should be set to ’full’),
NumHidden for MLP, and the Nu-KernelParam pair for SVM.

8.2 Classify General Regions using OCR

Sometimes it may be convenient to use the operators provided for OCR also for the classification of general objects.
This is possible as long as the objects can be described by the features that are provided for OCR (see section 7.7
on page 89). Note that many of the provided features are not rotation invariant. That is, if your objects have
different orientations in the images, you have to apply an alignment before applying the classification. The ex-
ample %HALCONEXAMPLES%\solution_guide\classification\classify_metal_parts_ocr.hdev shows
how to use OCR to classify the metal parts that were already classified with a general classification in the ex-
ample program %HALCONEXAMPLES%\solution_guide\classification\classify_metal_parts.hdev in
section 2 on page 11.
The program starts with the creation of an OCR classifier using create_ocr_class_mlp. There, the approx-
imated dimensions of the regions that represent the objects are specified by the parameters WidthCharacter
and HeightCharacter. The parameter Features is set to ’moments_central’. In contrast to the gen-
eral classification, no feature vectors have to be explicitly calculated and stored. This may enhance the speed
of the training as well as of the actual classification, and by the way needs less programming effort. The
parameter Characters contains a tuple of strings that defines the available class names, in this case the
classes ’circle’, ’hexagon’, and ’polygon’ are available. In %HALCONEXAMPLES%\solution_guide\
classification\classify_metal_parts.hdev the classes were addressed simply by their index, i.e., 0, 1,
and 2. There, the assignment of names for each class would have been possible, too, but then an additional tuple
with names must have been assigned and the correspondence between the class index and the class name must
have been made explicit.

create_ocr_class_mlp (110, 110, 'constant', 'moments_central', ['circle', \

'hexagon', 'polygon'], 10, 'normalization', 10, 42, \
OCRHandle)

Now, the input images, which are the same as already illustrated in figure 2.1 on page 12, and the class names for
the objects of each image are defined (FileNames and ClassNamesImage).

FileNames := ['nuts_01', 'nuts_02', 'nuts_03', 'washers_01', 'washers_02', \

'washers_03', 'retainers_01', 'retainers_02', \
'retainers_03']
ClassNamesImage := ['hexagon', 'hexagon', 'hexagon', 'circle', 'circle', \
'circle', 'polygon', 'polygon', 'polygon']

Then, the individual training regions of the objects are extracted from the training images. The procedure
to segment the regions is the same as used for %HALCONEXAMPLES%\solution_guide\classification\
classify_metal_parts.hdev in section 2 on page 11. The first region and its corresponding class name is
stored into an OCR training file using write_ocr_trainf. All following regions and their class names are stored
into the same training file by appending them via append_ocr_trainf.
 8.2 Classify General Regions using OCR D-95

for J := 0 to |FileNames| - 1 by 1
read_image (Image, 'rings/' + FileNames[J])
segment (Image, Objects)
count_obj (Objects, NumberObjects)
for k := 1 to NumberObjects by 1
select_obj (Objects, ObjectSelected, k)
if (J == 0 and k == 1)
write_ocr_trainf (ObjectSelected, Image, ClassNamesImage[J], \
'train_metal_parts_ocr.trf')
else
append_ocr_trainf (ObjectSelected, Image, ClassNamesImage[J], \

Tips
'train_metal_parts_ocr.trf')
endif
endfor
endfor

After adding all training samples to the training file, the training file is used by trainf_ocr_class_mlp to train
the ’font’, which here consists of three different shapes.

trainf_ocr_class_mlp (OCRHandle, 'train_metal_parts_ocr.trf', 200, 1, 0.01, \

Error1, ErrorLog1)

The images with the objects to classify are read in a loop and for each image the regions that represent the ob-
jects are extracted using the same procedure that was used for the training. Now, each region is classified using
do_ocr_single_class_mlp. Dependent on the classification result, the regions are visualized by different colors
(see figure 8.2).

Figure 8.2: Classifying metal parts because of their shape using the OCR specific classification operators: (left)
image with metal parts, (right) metal parts classified into three classes (illustrated by different gray
values).
D-96 General Tips

for J := 1 to 4 by 1
read_image (Image, 'rings/mixed_' + J$'02d')
segment (Image, Objects)
for k := 1 to NumberObjects by 1
select_obj (Objects, ObjectSelected, k)
do_ocr_single_class_mlp (ObjectSelected, Image, OCRHandle, 1, Class, \
Confidence)
if (Class == 'circle')
dev_set_color ('blue')
endif
if (Class == 'hexagon')
dev_set_color ('coral')
endif
if (Class == 'polygon')
dev_set_color ('green')
endif
dev_display (ObjectSelected)
endfor
endfor

8.3 Visualize the Feature Space (2D and 3D)

Sometimes, it may be suitable to have a look at the feature space, e.g., to check if the selected features build
clearly separable clusters. If not, another set of features should be preferred or further features should be added.
A reasonable visualization is possible only for the 2D (section 8.3.1) and 3D feature space (section 8.3.2), i.e.,
feature vectors or parts of feature vectors that contain only two to three features.

8.3.1 Visualize the 2D Feature Space

In section 3 on page 15, the example %HALCONEXAMPLES%\solution_guide\classification\

classify_citrus_fruits.hdev was coarsely introduced to explain what a feature space is. The example
classifies citrus fruits into the classes ’oranges’ and ’lemons’ and visualizes the 2D feature space for the
training samples. This feature space is built by the two shape features ’area’ and ’circularity’. In the
following, we summarize the steps of the example with the focus on how to visualize the 2D feature space.
At the beginning of the program, the names of the classes are defined and a GMM classifier is created. Then, inside
a for-loop the training images are read, the regions of the contained fruits are segmented from the red channel of
the color image (inside the procedure get_regions) and the features ’area’ and ’circularity’ are calculated
for each region (inside the procedure get_features). The values for the area of the regions are integer values.
As the feature vector has to consist of real values, the feature vector is converted into a tuple of real values before
it is added to the classifier together with the corresponding known class ID.
 8.3 Visualize the Feature Space (2D and 3D) D-97

ClassName := ['orange', 'lemon']

create_class_gmm (2, 2, 1, 'spherical', 'normalization', 10, 42, GMMHandle)
for I := 1 to 4 by 1
read_image (Image, 'color/citrus_fruits_' + I$'.2d')
get_regions (Image, SelectedRegions)
count_obj (SelectedRegions, NumberObjects)
for J := 1 to NumberObjects by 1
select_obj (SelectedRegions, ObjectSelected, J)
get_features (ObjectSelected, WindowHandle, Circularity, Area, \
RowRegionCenter, ColumnRegionCenter)
FeaturesArea := [FeaturesArea,Area]

Tips
FeaturesCircularity := [FeaturesCircularity,Circularity]
FeatureVector := real([Circularity,Area])
if (I <= 2)
add_sample_class_gmm (GMMHandle, FeatureVector, 0, 0)
else
add_sample_class_gmm (GMMHandle, FeatureVector, 1, 0)
endif
endfor
endfor

Now, the feature space for the oranges (dim gray) and lemons (light gray) of the training samples is visualized by
the procedure visualize_2D_feature_space (see figure 8.3).

visualize_2D_feature_space (Cross, Height, Width, WindowHandle, \

FeaturesArea[0:5], FeaturesCircularity[0:5], \
'dim gray', 18)
visualize_2D_feature_space (Cross, Height, Width, WindowHandle, \
FeaturesArea[6:11], FeaturesCircularity[6:11], \
'light gray', 18)

Figure 8.3: The feature space for the oranges (black) and lemons (gray) of the training samples.

Inside the procedure, first a 2D graph is created, i.e., depending on the width and height of the window, the
origin of the 2D graph is defined in image coordinates (OriginOfGraph), each axis of the graph is visualized by
an arrow (disp_arrow), and each axis is labeled with the name of the corresponding feature (set_tposition,
write_string).
D-98 General Tips

procedure visualize_2D_feature_space (Cross, Height, Width, WindowID,

FeaturesA, FeaturesC,
ColorFeatureVector, CrossSize):::
dev_set_color ('black')
OriginOfGraph := [Height - 0.1 * Height,0.1 * Width]
disp_arrow (WindowID, OriginOfGraph[0], OriginOfGraph[1], OriginOfGraph[0], \
Width - 0.2 * Width, 2)
disp_arrow (WindowID, OriginOfGraph[0], OriginOfGraph[1], 0.1 * Height, \
OriginOfGraph[1], 2)
set_tposition (WindowID, OriginOfGraph[0], Width - 0.2 * Width)
write_string (WindowID, 'Area')
set_tposition (WindowID, 0.07 * Height, OriginOfGraph[1])
write_string (WindowID, 'Circularity')

Then, the procedure determines the relations between the image coordinates and the feature values. For that, the
extent of the graph is defined on one hand in pixels for the image coordinate system (ExtentOfGraph) and on the
other hand in feature value units (RangeC, RangeA). Inside the image coordinate system, the extent is the same
for both axes and depends on the window height. For the feature values, the extent is defined for each feature axis
individually so that it covers the whole range of the corresponding feature values that is expected for the given set
of feature vectors. That is, for each feature, the extent corresponds to the approximated difference between the
expected maximum and minimum feature value. Having the extent of the graph in image coordinates as well as
the individual value ranges for the features, the scaling factor for each axis between feature values and the image
coordinate system is known (ScaleC, ScaleA).

ExtentOfGraph := Height - 0.3 * Height

RangeC := 0.5
RangeA := 24000
ScaleC := ExtentOfGraph / RangeC
ScaleA := ExtentOfGraph / RangeA

In addition to the scaling, a translation of the feature vectors is needed. Otherwise, the points representing the
feature vectors would be outside of the window. Here, the feature vectors are moved such that the position that is
built by the expected minimum feature values corresponds to the origin of the 2D graph in image coordinates. The
feature values at the origin are then described by MinC and MinA.

MinC := 0.5
MinA := 20000

In figure 8.4 the relations between the image coordinates and the feature values are illustrated.
Knowing the relations between the feature values and the image coordinate system, the procedure calculates the
image coordinates for each individual feature vector (RowFeature, ColumnFeature). For that, the distance of
each feature value to the origin of the 2D graph is calculated (in feature value units) and multiplied with the scaling
factor. The obtained distance in pixels (DiffC, DiffA) then simply is subtracted from respectively added to the
corresponding image coordinates of the graph’s origin (OriginOfGraph[0], OriginOfGraph[1]). The resulting
image coordinates are visualized by a cross contour that is created with the operator gen_cross_contour_xld
and displayed with dev_display.

NumberFeatureVectors := |FeaturesA|
for I := 0 to NumberFeatureVectors - 1 by 1
DiffC := ScaleC * (FeaturesC[I] - MinC)
DiffA := ScaleA * (FeaturesA[I] - MinA)
RowFeature := OriginOfGraph[0] - DiffC
ColumnFeature := OriginOfGraph[1] + DiffA
gen_cross_contour_xld (Cross, RowFeature, ColumnFeature, CrossSize, \
0.785398)
dev_display (Cross)
endfor
return ()

After visualizing the feature space for the training samples with the procedure visualize_2D_feature_space,
the training and classification is applied by the approach described in more detail in the sections of section 5 on
 8.3 Visualize the Feature Space (2D and 3D) D-99

OriginOfGraph[1]
Origin of
Image
Column
Circularity

MaxC

Feature Vector i
C_i X Row i = OriginOfGraph[0] − DiffC
OriginOfGraph[0]

Tips
DiffC=(ExtentOfGraph/RangeC)*(C_i−MinC)
<=> ExtentOfGraph

Column i = OriginOfGraph[1] + DiffA

DiffA=(ExtentOfGraph/RangeA)*(A_i−MinA)
RangeC

x
MinC Area
MinA A_i MaxA
RangeA
Row <=> ExtentOfGraph

Figure 8.4: Relations between image coordinates (gray) and feature values (black).

page 31.

train_class_gmm (GMMHandle, 100, 0.001, 'training', 0.0001, Centers, Iter)

for I := 1 to 15 by 1
read_image (Image, 'color/citrus_fruits_' + I$'.2d')
get_regions (Image, SelectedRegions)
count_obj (SelectedRegions, NumberObjects)
for J := 1 to NumberObjects by 1
select_obj (SelectedRegions, ObjectSelected, J)
get_features (ObjectSelected, WindowHandle, Circularity, Area, \
RowRegionCenter, ColumnRegionCenter)
FeaturesArea := [FeaturesArea,Area]
FeaturesCircularity := [FeaturesCircularity,Circularity]
FeatureVector := real([Circularity,Area])
classify_class_gmm (GMMHandle, FeatureVector, 1, ClassID, ClassProb, \
Density, KSigmaProb)
endfor
endfor

8.3.2 Visualize the 3D Feature Space

The example %HALCONEXAMPLES%\solution_guide\classification\visualize_3d_feature_space.hdev

shows how to visualize a 3D feature space for the pixels of two regions that contain differently texturized patterns.
The feature vector for each pixel is built by three gray values.
First, the feature vectors, i.e., the three gray values for each pixel have to be derived. For that, a texture image
is created by applying different texture filters (texture_laws), which are combined with a linear smoothing
(mean_image), to the original image. As we do not exactly know which laws filters are suited best to separate
D-100 General Tips

Figure 8.5: The two texture classes that have to be classified are marked by rectangles.

the specific texture classes from each other, we construct six differently filtered images and combine them to a six
channel image (compose6).

set_system ('clip_region', 'false')

read_image (Image, 'combine')
get_part (WindowHandle, Row1, Column1, Row2, Column2)
texture_laws (Image, ImageTexture1, 'ee', 5, 7)
texture_laws (Image, ImageTexture2, 'ss', 2, 7)
texture_laws (Image, ImageTexture3, 'rr', 0, 7)
texture_laws (Image, ImageTexture4, 'ww', 0, 7)
texture_laws (Image, ImageTexture5, 'le', 7, 7)
texture_laws (Image, ImageTexture6, 'el', 7, 7)
mean_image (ImageTexture1, ImageMean1, 41, 41)
mean_image (ImageTexture2, ImageMean2, 41, 41)
mean_image (ImageTexture3, ImageMean3, 41, 41)
mean_image (ImageTexture4, ImageMean4, 41, 41)
mean_image (ImageTexture5, ImageMean5, 41, 41)
mean_image (ImageTexture6, ImageMean6, 41, 41)
compose6 (ImageMean1, ImageMean2, ImageMean3, ImageMean4, ImageMean5, \
ImageMean6, TextureImage)

To get uncorrelated images, i.e., to discard data with little information, and to save storage, the six-channel image
is transformed by a principal component analysis. The resulting transformed image is then input to the procedure
gen_sample_tuples.

principal_comp (TextureImage, PCAImage, InfoPerComp)

gen_sample_tuples (PCAImage, Rectangles, Sample1, Sample2, Sample3)

Within the procedure, the first three images of the transformed texture image, i.e, the three channels with the
largest information content, are accessed via access_channel. Then, inside the image, for the two texture classes
rectangles are generated (see figure 8.5) and all pixel coordinates within these rectangles are determined and stored
in the tuples RowsSample and ColsSample. For each pixel the gray values of the first three channels of the
transformed texture image are determined and stored in the tuples Sample1, Sample2, and Sample3.
 8.3 Visualize the Feature Space (2D and 3D) D-101

procedure gen_sample_tuples (PCAImage, Rectangles, Sample1, Sample2,

Sample3):::
gen_empty_obj (ClassSamples)
Sample1 := []
Sample2 := []
Sample3 := []
gen_empty_obj (Rectangles)
access_channel (PCAImage, Image1, 1)
access_channel (PCAImage, Image2, 2)
access_channel (PCAImage, Image3, 3)
ClassNum := 0

Tips
I := 0
for Row := 80 to 340 by 260
for Col := 40 to 460 by 460
gen_rectangle1 (ClassSample, Row, Col, Row + 60, Col + 60)
concat_obj (Rectangles, ClassSample, Rectangles)
RowsSample := []
ColsSample := []
for RSample := Row to Row + 60 by 1
for CSample := Col to Col + 60 by 1
RowsSample := [RowsSample,RSample]
ColsSample := [ColsSample,CSample]
endfor
endfor
get_grayval (Image1, RowsSample, ColsSample, Grayvals1)
get_grayval (Image2, RowsSample, ColsSample, Grayvals2)
get_grayval (Image3, RowsSample, ColsSample, Grayvals3)
Sample1 := [Grayvals1,Sample1]
Sample2 := [Grayvals2,Sample2]
Sample3 := [Grayvals3,Sample3]
endfor
endfor
return ()

The feature vectors that are built by the gray values of the three channels are now displayed by the procedure
visualize_3d. To show the feature space from different views, it is by default rotated in discrete steps around
the y axis (RotY). Furthermore, the view can be changed by dragging the mouse. Then, dependent on the position
of the mouse pointer in the graphics window, the feature space is additionally rotated around the x and z axes.

for j := 0 to 360 by 1
dev_set_check ('~give_error')
get_mposition (WindowHandle, Row, Column, Button)
dev_set_check ('give_error')
if (Button != [])
RotX := fmod(Row,360)
RotZ := fmod(Column,360)
else
RotX := 75
RotZ := 45
endif
RotY := j
visualize_3d (WindowHandle, Sample1, Sample2, Sample3, RotX, RotY, \
RotZ)
endfor

Within the procedure visualize_3d, similar to the visualization of 2D feature vectors described in section 8.3.1
on page 96, the minimum and maximum values for the three feature axes are determined. Then, the maximum
value range of the features is determined and is used to define the scale factor for the visualization. In contrast
to the example used for the visualization of 2D feature vectors, the same scale factor is used here for all feature
axes. This is because all features are of the same type (gray values), and thus the ranges are in the same order of
magnitude.
D-102 General Tips

Min1 := min(Sample1)
Max1 := max(Sample1)
Min2 := min(Sample2)
Max2 := max(Sample2)
Min3 := min(Sample3)
Max3 := max(Sample3)
MaxFeatureRange := max([Max1 - Min1,Max2 - Min2,Max3 - Min3])
Scale := 1. / MaxFeatureRange

After defining a value for the virtual z axis, a homogeneous transformation matrix is generated and trans-
formed so that the feature space of interest fits completely into the image and can be visualized under the view
specified before calling the procedure. The homogeneous transformation matrix is built using the operators
hom_mat3d_translate, hom_mat3d_scale, and hom_mat3d_rotate. The actual transformation of the fea-
ture vectors with the created transformation matrix is applied with affine_trans_point_3d. To project the 3D
points into the 2D image, the operator project_3d_point is used. Here, camera parameters are needed. These
are defined as follows: the focal length is set to 0.1 to simulate a candid camera. The distortion coefficient κ is
set to 0, because no distortions caused by the lens have to be modeled. The two scale factors correspond to the
horizontal and vertical distance between two cells of the sensor, and the image center point as well as the width
and height of the image are derived from the image size.

DistZ := 7
hom_mat3d_identity (HomMat3DIdentity)
hom_mat3d_translate (HomMat3DIdentity, -(Min1 + Max1) / 2, \
-(Min2 + Max2) / 2, -(Min3 + Max3) / 2 + DistZ, \
HomMat3DTranslate)
hom_mat3d_scale (HomMat3DTranslate, Scale, Scale, Scale, 0, 0, DistZ, \
HomMat3DScale)
hom_mat3d_rotate (HomMat3DScale, rad(RotX), 'x', 0, 0, DistZ, \
HomMat3DRotateX)
hom_mat3d_rotate (HomMat3DRotateX, rad(RotY), 'y', 0, 0, DistZ, \
HomMat3DRotateY)
hom_mat3d_rotate (HomMat3DRotateY, rad(RotZ), 'z', 0, 0, DistZ, \
HomMat3DRotateZ)
affine_trans_point_3d (HomMat3DRotateZ, Sample1, Sample2, Sample3, Qx, Qy, \
Qz)
gen_cam_par_area_scan_division (0.1, 0, 0.00005, 0.00005, 360, 240, 720, \
480, CamParam)
project_3d_point (Qx, Qy, Qz, CamParam, Row, Column)

The result of the projection is a row and column coordinate for each feature vector. At this position, a region point
is generated and displayed (see figure 8.6).

gen_region_points (Region, Row, Column)

set_part (WindowHandle, 0, 0, 479, 719)
set_system ('flush_graphic', 'false')
clear_window (WindowHandle)
disp_obj (Region, WindowHandle)
set_system ('flush_graphic', 'true')
 8.3 Visualize the Feature Space (2D and 3D) D-103

Tips
Figure 8.6: The feature space shows two clearly separated clusters for the two texture classes.
D-104 General Tips
 Index D-105

Index

add characters to optical character recognition OCR classifier (kNN), 86

(OCR) training file, 82, 85, 87 OCR classifier (MLP), 80
add image training sample (GMM), 68 OCR classifier (SVM), 83
add image training sample (kNN), 69 optical character recognition (OCR), 7
add image training sample (MLP), 67 optimize classification parameters, 93
add image training sample (SVM), 68
add training sample (GMM), 49 pixel classification
add training sample (kNN), 53 detailed description, 57
operators, 64

Index
add training sample (MLP), 39
add training sample (SVM), 45 pixel classification (Euclidean), 73
approximate trained classifier (SVM), 46 pixel classification (GMM), 57, 68
pixel classification (Hyperbox), 73
classification pixel classification (kNN), 57, 69
first example, 11 pixel classification (MLP), 57, 67
overview, 7 pixel classification (SVM), 57, 67
theoretical background, 15
classification for optical character recognition read symbol (CNN), 88
(OCR), 75 read symbol (kNN), 87
operators, 78 read symbol (MLP), 82
classifier (deep learning), 23 read symbol (SVM), 85
classifier (GMM), 20, 47
segment image with pixel classification (GMM), 69
classifier (Hyperbox), 16
segment image with pixel classification (kNN), 70
classifier (kNN), 22, 52
segment image with pixel classification (MLP), 67
classifier (MLP), 18, 37
segment image with pixel classification (SVM), 68
classifier (SVM), 19, 42
segmentation, 7
classify data (GMM), 52
select approach for classification, 27
classify data (kNN), 55
select classifier training samples
classify data (MLP), 41
guide, 29
classify data (SVM), 47
select features for classification, 28
classify regions with optical character recognition,
set classification parameters (kNN), 54
94
speed up classifier (GMM), 47
create classifier (GMM), 48
speed up classifier (kNN), 52
create classifier (kNN), 53
speed up classifier (MLP), 37
create classifier (MLP), 37
speed up classifier (SVM), 42
create classifier (SVM), 42
train classifier (GMM), 50
Euclidean classifier, 16
train classifier (kNN), 53
evaluate feature vector (GMM), 51
train classifier (MLP), 40
evaluate feature vector (MLP), 41
train classifier (SVM), 45
features for OCR classifier, 89 train optical character recognition (OCR) (kNN), 86,
87
general classification, 31 train optical character recognition (OCR) (MLP), 80,
operators, 34 82
train optical character recognition (OCR) (SVM), 84,
novelty detection (classification), 7 85
novelty detection (with GMM classifier), 52 two-channel pixel classification, 72
novelty detection (with kNN classifier), 55
novelty detection (with SVM classifier), 44 visualize classification feature space, 96

object recognition 2D, 7 write optical character recognition (OCR) training

OCR classifier (CNN), 88 file, 82, 85, 87
D-106 Index