CHAPTER – 1

Digital Image Fundamentals

1
“One picture is worth more than ten
thousand words”

Anonymous

2
CONTENTS

 Image Fundamentals
 A simple image model
 Fundamental steps in DIP
 Elements of visual perception
 Image sensing and acquisition
 Image Sampling and quantization
 Basic relationship between pixels

3
 HISTORY OF DIGITAL IMAGE PROCESSING

Early 1920s: One of the first

applications of digital imaging was in
the news-paper industry
 The Bartlane cable picture
transmission service
 Images were transferred by Early digital image
submarine cable between London
and New York
 Pictures were coded for cable
transfer and reconstructed at the
receiving end on a telegraph printer
7
 HISTORY OF DIP (CONT…)
Mid to late 1920s: Improvements to the
Bartlane system resulted in higher quality
images
New reproduction
processes based
on photographic
techniques
Increased number Improved
of tones in digital image
reproduced images Early 15 tone digital image

8
HISTORY OF DIP (CONT…)

1960s: Improvements in computing

technology and the onset of the space
race led to a surge of work in digital
image processing
1964: Computers used to
improve the quality of
images of the moon taken A picture of the moon
by the Ranger 7 probe taken by the Ranger 7
probe minutes before
Such techniques were used landing
in other space missions
including the Apollo landings
9
HISTORY OF DIP (CONT…)
1970s: Digital image
processing begins to be used
in medical applications
1979: Sir Godfrey N.
Hounsfield & Prof. Allan M.
Cormack share the Nobel
Prize in medicine for the
invention of tomography,
the technology behind
Typical head slice CAT image
Computerised Axial
Tomography (CAT) scans

11
HISTORY OF DIP (CONT…)
1980s - Today: The use of digital image processing
techniques has exploded and they are now used for all
kinds of tasks in all kinds of areas
 Image enhancement/restoration
 Artistic effects
 Medical visualisation
 Industrial inspection
 Law enforcement
 Human computer interfaces

10
 WHAT IS AN IMAGE ? - THE VECTOR REPRESENTATION

 Object-oriented representation
 Does not show information of individual pixel,
but information of an object (circle, line, square,
etc.)

Circle (100, 20, 20)

Line (xa1, ya1, xa2, ya2)
Line (xb1, yb1, xb2, yb2)
Line (xc1, yc1, xc2, yc2)
Line (xd1, yd1, xd2, yd2)
9
COMPARISON
 Bitmap  Vector
 Can represent images  Can only represent
with complex simple line drawings
variations in colors, (CAD), shapes,
shades, shapes. shadings, etc.
 Larger image size  Efficient
 Fixed resolution  Flexible
 Easier to implement  Difficult to implement

10
WHAT IS AN IMAGE? - THE BITMAP REPRESENTATION

 Also called “raster or pixel maps” representation

 An image is broken up into a grid

y
Gray level
x
pixel

Original picture Digital image

f(x, y) I[i, j] or I[x, y] 1
1
WHAT IS A DIGITAL IMAGE?
A digital image is a representation of a two-
dimensional image as a finite set of digital values,
called picture elements or pixels

12
 WHAT IS A DIGITAL IMAGE CONTD..

Pixel values typically represent gray levels, colours,

heights, opacities etc
Remember digitization implies that a digital image is an
approximation of a real scene

1 pixel

13
 WHAT IS A DIGITAL IMAGE CONTD..
Common image formats include:
 1 sample per point (B&W or Grayscale)
 3 samples per point (Red, Green, and Blue)
 4 samples per point (Red, Green, Blue, and “Alpha”, a.k.a.
Opacity)

For most of this course we will focus on grey-scale images 14

 WHAT IS A DIGITAL IMAGE PROCESSING

Digital image processing focuses on two major tasks

 Improvement of pictorial information for human

interpretation
 Processing of image data for storage, transmission and
representation for autonomous machine perception

Some argument about where image processing ends and

fields such as image analysis and computer vision start.
The continuum from image processing to computer vision
can be broken up into low-, mid- and high-level processes.
15
Low Level Mid Level High Level
Process Process Process

Input: Image Input: Image Input: Attributes

Output: Image Output: Attributes Output:
Understanding
Examples: Noise Examples: Object
removal, image recognition, Examples: Scene
sharpening segmentation understanding,
autonomous
navigation

16
WHY IMAGE PROCESSING?
 Application
 Fingerprint retrieval
 Automatic target recognition
 Industrial inspection
 Medical imaging
 and more …
 Can commercial software do all the work?

17
A SIMPLE IMAGE MODEL
 Image: a two-dimensional light-intensity function;
f(x, y) = i(x, y) r(x, y)
where,
illumination 0 < i(x, y) < ∞,
reflectance components 0 < r(x, y) < 1
- illumination: the amount of source light incident
on the scene being viewed
- reflectance: the amount of light reflected by the
objects in the scene
 The interval [Lmin, Lmax]
: called the gray scale → [0, L]
18
EXAMPLES SHOWING THAT PERCEIVED BRIGHTNESS IS
NOT A SIMPLE FUNCTION OF INTENSITY.

19
 EXAMPLE OF SIMULTANEOUS CONTRAST
: ALL THE SMALL SQUARES HAVE EXACTLY THE SAME INTENSITY, BUT THEY
APPEAR PROGRESSIVELY DARKER AS THE BACKGROUND BECOMES LIGHTER.

20
EXAMPLES: IMAGE ENHANCEMENT
One of the most common uses of DIP techniques: improve
quality, remove noise etc..

21
 EXAMPLES: THE HUBBLE TELESCOPE

• Launched in 1990 the Hubble

telescope can take images of
very distant objects.
• However, an incorrect mirror
made many of Hubble’s
images useless.
• Image processing
techniques were
used to fix this.
22
EXAMPLES: ARTISTIC EFFECTS

Artistic effects are used

to make images more
visually appealing, to
add special effects and
to make composite
images.

23
EXAMPLES: MEDICINE
Take slice from MRI scan of canine heart, and find
boundaries between types of tissue
 Image with gray levels representing tissue density
 Use a suitable filter to highlight edges

Edge Detection Image 24

Original MRI Image of a Dog Heart
EXAMPLES: GIS
Geographic Information Systems
 Digital image processing techniques are used extensively
to manipulate satellite imagery
 Terrain classification
 Meteorology

25
EXAMPLES: GIS (CONT…)

Night-Time Lights of the

World data set
 Global inventory of
human settlement
 Not hard to imagine
the kind of analysis
that might be done
using this data

26
EXAMPLES: INDUSTRIAL INSPECTION

• Human operators are

expensive, slow and
unreliable
• Make machines do the
job instead
• Industrial vision systems
are used in all kinds of
industries.
• Can we trust them?

27
EXAMPLES: PCB INSPECTION
Printed Circuit Board (PCB) inspection
 Machine inspection is used to determine that all
components are present and that all solder joints are
acceptable
 Both conventional imaging and x-ray imaging are used

28
EXAMPLES: LAW ENFORCEMENT

Image processing techniques

are used extensively by law
enforcers
 Number plate
recognition for speed
cameras/automated toll
systems
 Fingerprint recognition
 Enhancement of CCTV
images
29
EXAMPLES: HCI
Try to make human computer
interfaces more natural
 Face recognition
 Gesture recognition
Does anyone remember the
user interface from
“Minority Report”?
These tasks can be
extremely difficult

30
 Original image of
Histogram Equalization Photoshop “Auto Contrast”
Mars and its moon Photoshop
APDG = 42.1 TEN = 12.1 APDG = 28.4 TEN = 22.0 result APDG = 72.9 TEN = 20.8

31
GLG in RGB space
GLG-RGB
APDG = 76.3 TEN = 27.0 GLG-HSI
APDG = 78.6 TEN = 26.6

From Zhiyu Chen’s preliminary proposal defense, January 2009

GOALS OF IMAGE PROCESSING

 Image improvement
 Improving the visual appearance of images to a
human viewer
 Image analysis

 Preparing images for measurement of the features

and structures present

32
 WHAT TO LEARN?
Preprocessing – low level High-level IP
Image Improvement Image Analysis

Image Image
Enhancement Restoration Image
Segmentation

Image Image Image

Acquisition Compression Coding Representation
& Description

Morphological Wavelet Recognition &

Image Processing Analysis Interpretation

Knowledge Base 33
FUNDAMENTAL STEPS IN DIGITAL IMAGE PROCESSING

Morphologi
Image
cal
Restoration
Processing

Image Segmentati
Enhancement on

Image Object
Acquisition Recognition

Problem Domain Representation

& Description
Colour Image Image 34

Processing Compression
KEY STAGES IN DIGITAL IMAGE PROCESSING:
IMAGE AQUISITION

Morphologic
Image
al
Restoration
Processing

Image Segmentati
Enhancement on

Image Object
Acquisition Recognition

Representation
Problem Domain & Description
Colour Image Image 35

Processing Compression
KEY STAGES IN DIGITAL IMAGE PROCESSING:
IMAGE ENHANCEMENT

Image Morphological
Restoration Processing
Image
Enhanceme Segmentati
nt on

Image Object
Acquisition Recognition

Representation
Problem Domain & Description
Colour Image 36
Image Compression
Processing
KEY STAGES IN DIGITAL IMAGE PROCESSING:
IMAGE RESTORATION

Image Morphological
Restoration Processing
Image
Enhanceme Segmentati
nt on

Image Object
Acquisition Recognition

Representation
Problem Domain & Description
Colour Image Image 37

Processing Compression
KEY STAGES IN DIGITAL IMAGE PROCESSING:
MORPHOLOGICAL PROCESSING

Image Morphological
Restoration Processing

Image Segmentati
Enhancement on

Image Object
Acquisition Recognition

Representation
Problem Domain & Description
Colour Image Image 38

Processing Compression
KEY STAGES IN DIGITAL IMAGE PROCESSING:
SEGMENTATION

Image Morphological
Restoration Processing
Image
Enhanceme Segmentati
nt on

Image Object
Acquisition Recognition

Representation
Problem Domain & Description
Colour Image Image 39

Processing Compression
KEY STAGES IN DIGITAL IMAGE PROCESSING:
OBJECT RECOGNITION

Image Morphological
Restoration Processing
Image
Enhanceme Segmentati
nt on

Image Object
Acquisition Recognition

Representation
Problem Domain & Description
Colour Image Image 40

Processing Compression
KEY STAGES IN DIGITAL IMAGE PROCESSING:
REPRESENTATION & DESCRIPTION

Image Morphological
Restoration Processing
Image
Enhanceme Segmentati
nt on

Image Object
Acquisition Recognition

Representation
Problem Domain & Description
Colour Image Image 41

Processing Compression
KEY STAGES IN DIGITAL IMAGE PROCESSING:
IMAGE COMPRESSION

Image Morphological
Restoration Processing
Image
Enhanceme Segmentati
nt on

Image Object
Acquisition Recognition

Representation
Problem Domain & Description
Colour Image Image 42

Processing Compression
KEY STAGES IN DIGITAL IMAGE PROCESSING:
COLOUR IMAGE PROCESSING

Image Morphological
Restoration Processing
Image
Enhanceme Segmentati
nt on

Image Object
Acquisition Recognition

Representation
Problem Domain & Description
Colour Image Image 43

Processing Compression
 IMAGE ACQUISITION

 Video camera
 Infrared camera

 Range camera

 Line-scan camera

 Hyper spectral camera

 Omni-directional camera

 and more …

44
 ELEMENTS OF VISUAL PERCEPTION

Digital Image Processing is built on foundation of

 Mathematical and Probabilistic formulations

 Human Intuition & analysis plays a key role in

choosing new technique over other.

 Study of Human vision system is thus important.

45
STRUCTURE OF HUMAN EYE

46
ROD & CONE CELLS

47
DISTRIBUTION OF ROD & CONE CELLS

48
 IMAGE SENSING AND ACQUISITION

Three principal sensor arrangement:

1) Image acquisition using a single sensor

49
 IMAGE SENSING AND ACQUISITION
Image Acquisition using a single sensor:

50
 IMAGE SENSING AND ACQUISITION
2) Image Acquisition using sensor strips:

51
 IMAGE SENSING AND ACQUISITION
3) Image Acquisition using sensor arrays:

52
IMAGE SENSING AND ACQUISITION

53
IMAGE SENSING AND ACQUISITION

54
 IMAGE SENSING AND ACQUISITION
Simple Image formation model:

o Images are denoted by 2-dimensional function of the form f( x, y).

o f at spatial co-ordinates ( x, y) is positive, finite scalar quantity

determined by source of the image.
0 < f( x, y) < ∞
o f( x, y) is characterized by two components:

o illumination: amount of source illumination incident on scene.

denoted by i( x, y)
o reflectance : amount of illumination reflected by the objects in
scene. denoted by r( x, y)
Thus, f(x,y) = i(x,y) r(x,y) where, 0 < i(x,y) < ∞ …... (1)
& 0(total absorption)< r( x, y) < 1(total reflectance) …… (2)
55
 IMAGE SENSING AND ACQUISITION
Simple Image formation model:

o Let the gray level (intensity) of a monochromatic image at any co-

ordinate be denoted by:
l = f(x0, y0);
From equation (1) & (2) we get
Lmin ≤ l ≤ Lmax
where, Lmin should be positive & Lmax be finite.

We have, Lmin = imin rmin & Lmax = imax rmax

The interval [Lmin, Lmax] is called the gray (intensity) scale.

We shift this interval to interval [0, L-1],

where, l = 0 is black & l = L-1 is white on grey scale. 56
IMAGE SAMPLING & QUANTIZATION

57
IMAGE SAMPLING & QUANTIZATION

58
SAMPLING & QUANTIZATION

f (x, y): digitized both spatially and in

amplitude
Digitization of the spatial coordinates (x, y)
called image sampling
 Amplitude digitization

: called gray-level quantization

 Resolution: the degree of discernible detail of
an image depends strongly on the number of
samples and gray-levels.

59
SCANNING & SAMPLING
 Scanning

Progressive scan Interlaced scan

 Sampling

60

Case of interlaced scan

Case of progressive scan
SAMPLING & QUANTIZATION

 The more parameters (the number of samples and

gray levels) are increased, the closer the digitized
array approximates the original image.
 The digitization process requires decisions about
values for N, M, and the number of discrete gray
levels allowed for each pixel.

N = 2n, M = 2k and G = 2m
b=NMm
where, G: the number of gray levels
b: the number of bits required to store a
digitized image.
61
EFFECTS ACCORDING TO SPATIAL RESOLUTION &
NUMBER OF BITS

 Effects of reducing spatial resolution ; pixel

replication
 produced a checkerboard effect

 Effectsproduced by decreasing the number of bits

used to represent the number of gray levels in an
image.
 A 10241024 image displayed in 256, 128, 64, 32,
16, 8, 4, and 2 levels, respectively.

62
EFFECT OF REDUCING SPATIAL RESOLUTION:
(A) 10241024, 256-LEVEL DIGITAL IMAGE OF A ROSE
(B)~(F) RESULTS OF REDUCING THE SPATIAL RESOLUTION FROM N=1024 TO
N=512, 256, 128, 64, AND 32, RESPECTIVELY.

63
EFFECT OF DECREASING THE NUMBER OF BITS:
(A) 10241024, 256-LEVEL DIGITAL IMAGE OF A ROSE
(B)~(F) RESULTS OF REDUCING THE NUMBER OF BITS FROM M=7 TO M=1,
RESPECTIVELY.

64
SAMPLING & QUANTIZATION (3)
 Isopreference curve correspond to images of equal
subjective quality
- the quality of the images tends to increase as N
and m are increased.
- a decrease in m generally increases the
apparent contrast of an image.
 for images with a larger amount of detail only
a few gray levels are needed.

65
ISOPREFERENCE CURVES

66
SAMPLING & QUANTIZATION (4)

 Non-uniform sampling and quantization

: depends on the characteristics of the image

- Fine sampling is required in the neighborhood of

sharp gray-level transitions, whereas coarse sampling
may be utilized in relatively smooth regions.

- When the number of gray levels must be kept small,

the use of unequally spaced levels in the quantization
process usually is desirable ( called tapered
quantization).

67
 IMAGE SAMPLING & QUANTIZATION

Spatial & Intensity Resolution

Spatial Resolution:
dpi : (dots per inch)

Newspapers : 75 dpi
Magazines : 133 dpi
Glossy brochures : 175 dpi
Books : 2400dpi

68
 IMAGE SAMPLING & QUANTIZATION

Intensity resolution: smallest discernible change in intensity level.

Image interpolation: process of using known data to estimate the

values at unknown locations.

 Nearest neighbor interpolation: assigns intensity of nearest

neighbor in original image.

 Bilinear interpolation: uses 4 nearest neighbors.

 Bicubic interpolation: uses 16 nearest neighbors.

69
 HIGH DEFINITION (HD)

Video Frame size in Pixels per

Scanning type Frame rate (Hz)
mode pixels (W×H) image

23.976, 24, 25, 29.97,

720p 1,280x720 921,600 Progressive
30, 50, 59.94, 60, 72
25 (50 fields/s), 29.97
1080i 1,920x1,080 1,036,800 Interlaced (59.94 fields/s), 30 (60
fields/s)
23.976, 24, 25, 29.97,
1080p 1,920x1,080 2,073,600 Progressive
30, 50, 59.94, 60

70
HIGH DEFINITION (HD)

72
BASIC RELATIONSHIPS BETWEEN PIXELS

• Neighborhood
• Adjacency
• Connectivity
• Paths
• Regions and boundaries

Neighbors of a Pixel
• Any pixel p(x, y) has two vertical and two horizontal neighbors, given by
(x+1, y), (x-1, y), (x, y+1), (x, y-1)
• This set of pixels are called the 4-neighbors of P, and is denoted by N4(P).
• Each of them are at a unit distance from P.
The four diagonal neighbors of p(x,y) are given by,
(x+1, y+1), (x+1, y-1), (x-1, y+1), (x-1 ,y-1)
• This set is denoted by ND(P).
• Each of them are at Euclidean distance of 1.414 from P.
8-neighbors of a pixel p are its vertical horizontal and 4 diagonal neighbors denoted by73
N8(p)
ADJACENCY

Let V be set of gray levels values used to define adjacency.

• 4-adjacency: Two pixels p and q with values from V are 4-
adjacent if q is in the set N4(p).
• 8-adjacency: Two pixels p and q with values from V are 8-
adjacent if q is in the set N8(p).
• m-adjacency: Two pixels p and q with values from V are
madjacent
if,
– q is in N4(P).
– q is in ND(p) and the set [ ] is empty
(has no pixels whose values are from V).

74
CONNECTIVITY :
4-connected, if q is in the set N4(p)
b. 8-connected, if q is in the set N8(p)
c. m-connected, if
i. q is in N4(p) or
ii. q is in ND(p) and the set
is empty

Paths & Path lengths

A path from pixel p with coordinates

(x, y) to pixel q with coordinates (s, t)
is a sequence of distinct pixels with
coordinates: (x0, y0), (x1, y1), (x2, y2) …
(xn, yn), Here n is the length of the
path. 75
DISTANCE MEASURES
Given pixels p, q and z with coordinates (x, y), (s, t), (u, v)
respectively, the distance function D has following properties:
a. D(p, q) ≥0 , [D(p, q) = 0, iff p = q]
b. D(p, q) = D(q, p)
c. D(p, z) ≤D(p, q) + D(q, z)

The following are the different Distance

measures:
a. Euclidean Distance :
De(p, q) = [(x-s)2 + (y-t)2]
b. City Block Distance:
D4(p, q) = |x-s| + |y-t|
c. Chess Board Distance:
D8(p, q) = max(|x-s|, |y-t|)
76
DISCUSSION (Q&A)

77