Deep Learning-Based Approaches for Contactless

Fingerprints Segmentation and Extraction

† § §
M.G. Sarwar Murshed Syed Konain Abbas Sandip Purnapatra
§ §
Daqing Hou Faraz Hussain
arXiv:2311.15163v1 [[Link]] 26 Nov 2023

†
University of Wisconsin–Green Bay, WI, USA
{murshedm}@[Link]
§
Clarkson University, Potsdam, NY, USA
{abbas,purnaps,dhou,fhussain}@[Link]

Abstract
Fingerprints are widely recognized as one of the most unique and reliable charac-
teristics of human identity. Most modern fingerprint authentication systems rely on
contact-based fingerprints, which require the use of fingerprint scanners or fingerprint
sensors for capturing fingerprints during the authentication process. Various types of
fingerprint sensors, such as optical, capacitive, and ultrasonic sensors, employ distinct
techniques to gather and analyze fingerprint data. This dependency on specific hard-
ware or sensors creates a barrier or challenge for the broader adoption of fingerprint
based biometric systems. This limitation hinders the widespread adoption of finger-
print authentication in various applications and scenarios. Border control, healthcare
systems, educational institutions, financial transactions, and airport security face chal-
lenges when fingerprint sensors are not universally available.
To mitigate the dependence on additional hardware, the use of contactless fin-
gerprints has emerged as an alternative. Developing precise fingerprint segmentation
methods, accurate fingerprint extraction tools, and reliable fingerprint matchers are
crucial for the successful implementation of a robust contactless fingerprint authenti-
cation system. This paper focuses on the development of a deep learning-based seg-
mentation tool for contactless fingerprint localization and segmentation. Our system
leverages deep learning techniques to achieve high segmentation accuracy and reli-
able extraction of fingerprints from contactless fingerprint images. In our evaluation,
our segmentation method demonstrated an average mean absolute error (MAE) of 30
pixels, an error in angle prediction (EAP) of 5.92 degrees, and a labeling accuracy of
97.46%. These results demonstrate the effectiveness of our novel contactless fingerprint
segmentation and extraction tools.

Keywords— Contactless, Deep learning, Segmentation, Fingerprints, Biometrics

1
1 Introduction
Biometric recognition systems play important roles in different domains due to their accu-
racy and quick processing time. Biometric authentication is based on an individual’s distinct
physiological or behavioral traits, such as voice prints, iris patterns, or fingerprints [17]. It
reduces the risk of unauthorized access and burdens like the requirement of memorizing
passwords, fraudulent authentication, and password stealing, as most of the biometric traits
are difficult to forge or replicate, providing a higher level of security compared to passwords,
which can be easily guessed or stolen unless the password is strong. Therefore, such authen-
tication systems are becoming increasingly common in situations where users need to quickly
authenticate themselves, such as border crossings, airports, cell-phone based authentication,
and in healthcare, by simply providing their biometric data [16]. This eliminates the need
for complex login procedures or repetitive authentication steps.
One of the most popular and extensively utilized biometric characteristics for authentica-
tion is fingerprints. Each fingerprint is unique and remains consistent over time. To enhance
security and accuracy, multiple fingerprints, such as the fingerprint slap, are used for a user
instead of a single fingerprint. The first step in using a slap fingerprint to authenticate a user
is to separate or segment each finger in the slap [2]. There are several publicly and commer-
cially available slap fingerprints segmenters, such as NIST NFSEG [10] and Neurotechnology
Verifinger Segmenter [5].
We have developed a deep learning-based fingerprint segmenter called CRFSEG [15]
which outperforms other fingerprint segmenters in terms of precise fingerprint detection
and fingerprint matching accuracy. However, it is important to note that these fingerprint
segmenters have mainly been developed and used for contact-based slap fingerprint images.
While contact-based slap fingerprint authentication offers significant advantages, it also has
multiple limitations, including the requirement for specialized hardware or infrastructure for
implementation.
Most fingerprint authentication systems require physical contact between the fingers and
fingerprint-capturing devices to capture fingerprint images. However, such contact can re-
sult in problems like blurred images due to dust on the capturing surface of the device and
the potential for contamination from previously captured fingerprint information. Further-
more, fingerprint data can raise privacy concerns, and there is a possibility of false positives
or false negatives during the matching process. In addition to these limitations, contact-
based fingerprint-capturing processes raise health concerns, particularly during pandemics,
as they require direct contact with capturing devices [18]. However, ongoing advancements
in fingerprint-based biometric technology continue to address these challenges, making bio-
metrics a promising and increasingly adopted authentication solution. One promising area
is the use of contactless fingerprinting, which eliminates the need for additional specialized
hardware or infrastructure. The term fingerphoto refers to a contactless fingerprint image
that may be taken with any camera, including a simple mobile phone camera. A fingerphoto
is created by capturing an image of human fingers with a basic smartphone camera and
typically includes multiple fingers [14]. In Figure 1, a sample fingerphoto is shown.
Contactless fingerprint-based authentication systems offer several advantages over tradi-
tional contact-based fingerprint authentication systems:

2
Figure 1: A sample fingerphoto image was taken with a simple mobile phone camera. Fin-
gerphotos are typically acquired by capturing an image of human fingers using a standard
smartphone camera, and they frequently include multiple fingers within the frame.

i) Improved security against spoofing: One of the most important advantages of using
contactless fingerprint biometric systems is that they often incorporate additional anti-
spoofing measures to enhance security. Advanced sensors and algorithms can detect
and differentiate between live fingers and spoofing attempts using fake or synthetic
fingerprints, reducing the risk of unauthorized access.
ii) Hygiene and cleanliness: By eliminating the requirement of physical contact between
the user’s finger and the fingerprint-capturing sensor, contactless fingerprint authen-
tication reduces the risk of spreading germs, viruses, and bacteria, making it a more
hygienic option, especially in high-traffic areas or shared devices.
iii) Convenience: Contactless fingerprint authentication is often faster and more user-
friendly than contact-based methods. Users can simply place their finger toward the
sensors/cameras or the camera near the user’s hand, eliminating the need for precise
alignment or direct physical contact. Most contactless fingerprint sensors like Idemia
MorphoWave, Fujitsu PalmSecure can be used from a distance, which can be helpful
for people with disabilities or who have difficulty bending over. This improves the
overall user experience and can lead to increased user adoption and satisfaction.
iv) Versatility and ease of integration: Contactless fingerprint biometric systems can be
integrated into a wide range of devices and environments. They can work with exist-
ing touchless technologies, such as proximity sensors or facial recognition, to provide
multi-modal authentication options. This versatility allows for seamless integration
into various applications, including access control systems, smartphones, and payment
terminals.

Despite the advantages offered by contactless fingerprint biometrics, it is worth noting

that they also have some limitations. These include potential susceptibility to environmental
factors such as lighting conditions, distance captures, and uneven focus, as well as the need
for appropriate sensor technology. The implementation of multispectral imaging technol-
ogy exemplifies a suitable approach to address certain challenges associated with contactless

3
fingerprint biometrics. Multispectral fingerprint sensors excel in overcoming environmental
obstacles by capturing data across multiple wavelengths of light enabling them to mitigate
issues arising from factors such as lighting conditions and other environmental variables.
Advancements in sensor technology and algorithmic improvements continue to help over-
come these challenges and make contactless fingerprint biometrics a promising and viable
authentication solution [1].
Most contactless fingerprint authentication systems utilize fingerphotos, which are con-
tactless fingerprint images that are captured from multiple fingers. Segmenting all fingertips
from fingerphotos is an active area of research in contactless biometric authentication sys-
tems. Segmentation plays a crucial role in fingerprint matching, as observed in the existing
literature [13].
In this paper, we describe our novel contactless fingerprint segmentation system devel-
oped by enhancing our existing contact-based fingerprint segmentation model, CRFSEG
(Clarkson Rotated Fingerprint Segmentation Model) [15], by updating its architecture and
training it on a contactless dataset. Our novel contactless segmentation model CRFSEG-v2
demonstrates higher accuracy when evaluated on our in-house contactless fingerprint dataset
consisting of 23,650 slaps, which were annotated by human experts. We describe a novel
model for segmentation of contactless fingerphotos. Its architecture is based on our prior
fingerprint segmentation system [15]. Special attention was given to optimizing the deep
learning architecture for this purpose. The paper presents the following novel contributions:

• Built an in-house contactless dataset that contains 23,650 fingerphotos (94,600 single
fingers).

• Annotated all fingerphoto images1 manually to establish a ground-truth baseline for

the accuracy assessment of fingerprint segmentation systems.

• Updated and retrained for contactless fingerphotos our previously developed age-invariant
deep learning-based slap segmentation model, that can handle arbitrarily orientated
fingerprints.

• Assessed the performance of the contactless model named CRFSEG-v2 (Clarkson Ro-
tated Fingerprint Segmentation Model) using the following metrics, MAE, EAP, and
accuracy in fingerprint matching.

2 Related Work
Researchers have utilized various segmentation methods to achieve precise contactless fin-
gerprint segmentation. These methods employ different algorithms, including convolutional
neural networks (CNN), recurrent adversarial learning, fuzzy C-mean (FCM), and genetic
algorithm (GA) [4]. Wild et al. introduced a Skin-Mask finger segmentation technique to
segment contactless finger photos [21]. The study employed a filtering approach utilizing
the COTS fingerprint matcher (Verifinger) and the NFIQ 2.0 quality metric. They used a
1
The total number of images is 23,650. Out of 23,650, 2150 fingerphotos were human annotated and
21,500 were generated by augmentation method.

4
dataset comprising 2582 contact-based and 1728 contactless images from 108 fingers, result-
ing in TAR values ranging from 95.5% to 98.6% at FAR=0.1%. It involves reducing the
image resolution, creating masks based on the HLS (Hue, Lightness, and Saturation) color
space, and selecting finger pixels based on correlation values. The technique also includes
morphological operations, variance calculations, and contour extraction to obtain individual
finger images. However, the skin-mask finger segmentation technique relies on the assump-
tion that the skin color of the fingers is relatively uniform. This assumption may not hold
true in all cases, such as when the fingers are dirty or have been exposed to sunlight. This
type of segmentation is sensitive to environmental factors such as lighting conditions and
variations in skin color. The presence of objects or background elements that resemble skin
tones may affect the accuracy of the segmentation results. Additionally, the technique may
not include all finger pixels in the selection, and the separation of fingers connected via the
thumb relies on evaluating local minima and maxima, which may not always be accurate.
Malhotra et al. introduced a method for segmenting the distal phalange in contactless
fingerprint images by combining a saliency map and a skin-color map [12]. They adopted
a different approach by employing a random decision forest matcher and feature extraction
with a deep scattering network. Their dataset included 1216 contact-based and 8512 con-
tactless images from 152 fingers, resulting in an EER ranging from 2.11% to 5.23%. This
process involves extracting a binary mask that represents the finger region in a captured
finger selfie. It combines region covariance-based saliency and skin color measurements for
effective segmentation. The steps include extracting visual features, constructing covariance
matrices, computing dissimilarities between regions, generating saliency maps, converting to
the CMYK (Cyan(C), Magenta(M), Yellow(Y), and BlacK(B)) color model, fusing saliency
and skin color maps, and applying thresholds to obtain the final segmented mask. Although
the algorithm produces impressive results, it requires extensive tuning of hyperparameters
and it continues to struggle with accurately distinguishing fingerprints in the presence of
noisy backgrounds or under excessively bright lighting conditions.
To address challenges in accurately segmenting fingerprints under challenging illumina-
tion conditions or noisy backgrounds, Grosz et al. an auto-encoder-based segmentation
approach [4]. To perform 500 PPI deformation and scale correction on contactless finger-
prints, they utilized a spatial transformer. Their dataset included three parts one with
8,512 contactless and 1,216 contact-based fingerprints from 152 fingers, another containing
2,000 contactless and 4,000 contact-based fingerprints from 1,000 fingers, and a ZJU dataset
comprising 9,888 contactless and 9,888 contact-based fingerprints from 824 fingers. These
datasets served as the basis for evaluating their methodology. Their approach achieved im-
pressive EERs of 1.20%, 0.72%, 0.30%, and 0.62% on these datasets, respectively. A U-net
segmentation network is employed to segment a distal phalange of contactless fingerprint
photos. The segmentation network is trained using manually marked segmentation masks
from a dataset. The segmentation algorithm takes unsegmented images as input and out-
puts a segmentation mask. This mask is then used to crop a distal phalange of fingerprints
and the background is removed. Image enhancements, including histogram equalization and
gray-level inversion, are applied to improve ridge-valley structure.
None of the aforementioned research studies provide information about the accuracy like
mean absolute error (MAE), error in angle prediction (EAP), and labeling accuracy of their
finger photo segmentation techniques. Our work addresses this gap by evaluating our novel

5
contactless segmentation system using a large dataset of images. We not only report the
accuracy of the segmenter but also assess its impact on contactless fingerprint-matching
accuracy.

3 Research Methods
In this section, we provide a detailed examination of the method used for collecting finger-
photos from adult subjects, data annotation, data augmentation, and ground truth labeling.
Then, we describe the proposed neural network architecture of CRFSEG-v2. Finally, we
discuss the metrics used to evaluate different slap segmentation algorithms.

3.1 Slap Dataset

In this section, we discuss contactless datasets including fingerphoto collection, annotation,
augmentation, and ground truth creation.

3.1.1 Contactless Fingerphoto Collection

Contactless fingerphoto collection is a crucial component of contactless fingerprint biomet-
rics research. Our literature review revealed a shortage of contactless fingerprint images,
particularly finger photos containing multiple fingers. To address this gap, we conducted
a user study and collected fingerphotos multiple times. The collection process involved ac-
quiring fingerprint images without any physical contact with fingerprint capture devices. In
this subsection, we discuss the different techniques and considerations involved in contactless
fingerphoto collection.

• Camera Selection: The selection of a camera is of utmost importance in contactless

finger photo collection. Opting for high-resolution cameras with excellent color repro-
duction and minimal noise levels is ideal for capturing clear and detailed finger images.
However, when developing a contactless authentication system that operates in real-
world scenarios, relying solely on high-quality images is not recommended. Therefore,
we utilized multiple mobile devices, including the Google Pixel 2, iPhone 7, X, Samsung
S6, S7, and S9 to capture fingerphotos.

• Illumination Conditions: The lighting conditions during the capture of contactless

fingerphotos are crucial. It is important to have sufficient and uniform illumination
to minimize shadows, reflections, and image noise. We took precautions to avoid
overexposure or underexposure, as these can affect the visibility of fingerprint details.

• Finger Alignment and Placement: Providing clear instructions and visual cues can
help users achieve optimal finger placement. However, for real-world applications,
fingerprint alignment and placement can vary. Therefore, we did not instruct users on
how to position their fingers specifically within the camera’s field of view. As a result,
we obtained fingerprints with diverse alignment and placement.

6
 • Motion Blur Reduction: Movement during image capture can cause motion blur, which
can degrade the quality of finger images. To minimize motion blur and improve image
sharpness, users were instructed to place their hands on a table or stable surface.

• Backgrounds: Selecting an appropriate background is crucial for improving the contrast

between the finger and the background, which facilitates accurate extraction of the
finger region. However, it is not always possible for users to have an ideal background in
real-world applications. Therefore, we employed various backgrounds when capturing
fingerphotos.

• Privacy and Data Protection: Collecting contactless fingerphotos involves capturing

personal biometric data. It is crucial to comply with privacy regulations and en-
sure the secure storage and handling of the collected data. To protect the privacy
and confidentiality of individuals’ biometric information, we employed anonymization
techniques, and with appropriate permits from the Institutional Review Board (IRB).

This dataset comprises a total of 2150 fingerphotos. A comprehensive and diverse image
dataset should encompass a broad range of scenarios, including different poses, illumination
conditions, sizes, brightness levels, and fingerprint positions. Such datasets are valuable for
developing a robust deep-learning-based fingerprint segmentation model and are compiled
to thoroughly evaluate the model’s performance in real-world scenarios. To introduce such
variation into our dataset, we employed data augmentation techniques. Through augmen-
tation, we obtained an additional 21,500 augmented images, thereby expanding the dataset
to a total 23,650 images.

3.2 Fingerphoto Annotation and Augmentation

Data annotation plays a crucial role in developing a well-structured and representative
dataset, which is essential for building a reliable deep learning-based image segmentation
model. The annotation process involves delineating a bounding box around each fingertip
in a fingerphoto and assigning a corresponding label to each fingertip. The fingertip la-
bels used include Left-Index, Left-Middle, Left-Ring, Left-Little, Left-Thumb, Right-Index,
Right-Middle, Right-Ring, Right-Little, and Right-Thumb. This annotated data is then uti-
lized to train a deep learning model, enabling it to recognize and classify distinct objects or
classes within a fingerphoto.
The creation of a ground truth dataset for evaluating contactless fingerprint segmenta-
tion models typically involves laborious manual annotation, especially when dealing with
a large number of images. To deal with this process, we initially annotated 200 images
and fine-tuned our previously developed contact-based fingerprint segmentation model using
this annotated dataset. Subsequently, we employed the fine-tuned CRFSEG-v2 segmenta-
tion model to automatically segment the fingerphotos. However, a manual review of the
CRFSEG-v2 results was still necessary to ensure accuracy and alignment. This involved
visually examining the images, rectifying errors, and aligning the fingerprints. Although the
CRFSEG-v2 model demonstrated proficiency in generating bounding boxes around finger-
tips, it occasionally mislabeled them. We rectified these misclassifications through manual

7
inspection of all the fingerphotos. This automated segmentation approach resulted in a 65-
70% reduction in the time required for annotation. We further augmented the contactless
dataset by rotating all the fingerphotos at various angles (-90° to 90°) to create a diverse set
of slap images containing rotated fingerprints. The details of the dataset are shown in the
Table 1.

Table 1: We have 23,650 finger photos. Out of these, 2,150 were collected by Clarkson
University. For image collection, we used different types of mobile phones such as Samsung
S20, iPhone 7, iPhone X, and Google Pixel. Subsequently, we annotated the images manually.
To create rotating slap images, we utilized the 2,150 finger photos that were annotated
by humans to create more images using an augmentation technique. This resulted in the
generation of 21,500 more augmented images by rotating all the finger photos at different
angles (-90 to 90 degrees). This augmentation is intended to help the model become invariant
to different types of rotations of finger photos.
Dataset Total fingerphotos Lefthand fingerphotos Righthand fingerphotos
Bonafide 2150 1118 1032
Augmented 21500 11180 10320
Total 23650 12298 11352

3.3 Deep Learning Architecture for Contactless fingerphoto Seg-

mentation
In prior work, we developed a two-stage Faster R-CNN architecture for segmenting contact-
based slap fingerprints [15]. We utilized a similar two-stage faster architecture for seg-
menting contactless slap fingerphotos. However, contact-based and contactless fingerprint
images possess distinct characteristics in terms of quality, structure, size, and other factors.
Contact-based images tend to exhibit higher quality as they are captured using physical
sensors that make direct contact with the finger. Moreover, these images benefit from stable
lighting conditions. In contrast, contactless images often suffer from challenges such as blur-
riness, noise, diverse backgrounds, and variations in illumination, as they are captured from
a distance using different types of sensors. Consequently, we needed to make small modifi-
cations to the architecture to accommodate the lower quality and address other difficulties
inherent in contactless images. This modification involves the use of different backbone net-
works, loss functions, increasing the kernel size, and adjusting the learning rate and number
of epochs. The contactless fingerphoto deep learning architecture, see Figure 2, comprises
three key structural components: the box head, the oriented region proposal network, and
the backbone network.

3.3.1 Backbone Network

CRFSEG-v2 backbone network, featuring the Resnet-100 with FPN architecture, is designed
to extract feature maps of different sizes from contactless fingerphotos. This network incorpo-
rates both stem blocks and bottleneck blocks. To optimize calculations, extract key features,
and expand channel capabilities, bottleneck blocks with three convolution layers of varying

8
kernel sizes (1×1, 3×3, and 1×1) are employed [7]. Within the stem block, the focus is on re-
ducing the input image size, achieved through Convolution 2D layers, ReLU activation, and
max-pooling layers, to minimize computational cost while retaining all essential information.
The oriented region proposal network (O-RPN) then takes in the multiscale semantic-rich
feature maps generated by the backbone network.

3.3.2 Oriented Region Proposal Network (O-RPN)

After being obtained from the backbone network, the input image feature maps are employed
by the Oriented Region Proposal Network (O-RPN) to generate proposals for oriented regions
of interest (ROIs), which are most likely to encompass the objects of interest. This is achieved
by identifying areas with a high probability of containing the target objects. The proposal of
ROIs by the O-RPN is a crucial step as it helps in generating accurate rotated bounding boxes
for both axis-aligned and rotated objects. This is in contrast to the conventional regional
proposal network (RPN) used in the Faster R-CNN architecture, where only axis-aligned
regions are proposed.
The proposal of oriented regions is achieved by using a sliding window approach with
a 3×3 kernel on the feature maps. Consequently, the system generates anchor boxes with
diverse aspect ratios, scales, and orientations. If we have ka different orientations, ks different
scales, and kr different aspect rations, then K = ka × ks × kr anchor boxes are generated for
each position in the feature map. Most of these anchor boxes might not have targeted objects
in them. Subsequently, convolution layers and parallel output layers called localizing and
classifying layers are employed to differentiate anchor boxes with target objects from others.
The classifying layer assigns foreground or background labels according to the intersection-
over-union (IoU) score with ground truth boxes while learning offsets (x,y,w,h, θ) is done
by the localization layer for the foreground boxes. The term "Foreground" is used to denote
regions of interest (proposals) that exhibit a substantial overlap with the bounding boxes
of the ground truth objects in the image. In simpler terms, these are areas most likely
to contain a prominent object. The classifying layer assigns a "foreground" label to these
regions. Conversely, regions of interest that lack a significant overlap with any ground truth
bounding box are designated as the "background". The probability of an object being present
in these areas is lower. The classifying layer assigns a "background" label to these regions.
Regarding the regression offset, the localizing layer generates (K×5) parameterized encod-
ings, while the classifying layer produces (K×5) parameterized scores for region classification.
The anchor generation strategy is illustrated in Figure 2.
For training O-RPN, seven different orientations (-π/4, -π/6, -π/12, 0, π/12, π/6, π/4),
three aspect ratios (1:1, 1:2, 2:1 ) and three scales (128, 256 and 512) are used to generate
anchors. In conclusion, the selection of seven orientations, three aspect ratios, and three
scales involves a trade-off between maintaining computational efficiency during training and
inference while ensuring the model’s capacity to capture the diversity of objects in images
[8]. Then, all the anchors are divided into three categories

• positive anchors: the Intersection over Union (IoU) overlap between these anchors
and the ground-truth boxes exceeds 0.7, indicating a strong alignment with the target
objects,

9
Figure 2: The complete architecture for CRFSEG-v2 includes several processing stages and
is intended for precise fingerprint segmentation. For the feature maps, we need to provide our
input contactless fingerprint image to a pre-trained CNN model on the ImageNet dataset.
To generate oriented anchors, O-RPN needs to run on all levels of feature maps. By selecting
spatial features from O-RPN and from the output of the backbone network, ROI pooling lay-
ers generate fixed-length feature vectors. These fixed-length feature vectors are then passed
through the fully connected layers. There are two parallel branches that receive the output
of the fully connected layers, referred to as the Softmax Classifier and Oriented Bounding
Box Regressor. The softmax layers contain a softmax layer for multiclass classification, and
the regressors contain bounding box regression.

• negative anchors: these anchors have an IoU overlap smaller than 0.3 with the ground-
truth boxes, indicating a significant mismatch with the target objects,

• neutral anchors: these anchors have an IoU overlap between 0.3 and 0.7 using the
ground-truth boxes. These are removed from the anchor set and not used during
subsequent processing.

In contrast to Faster R-CNN, where horizontal anchor boxes are used, our approach utilizes
oriented anchor boxes and oriented ground-truth boxes. The loss functions employed to train
the O-RPN are defined by the following equations:

Lo−rpn = Lcls (p, u) + λuLreg (t, t∗ ) (1)

Here, Lcls is the classification loss, p is the predicted probability across the foreground and
background classes by the softmax function, u represents class label for anchors, where u
= 1 for foreground containing fingerprint and u = 0 for background; t = (tx , ty , th , tw , tθ )
denotes the predicted regression offset value of an anchor calculated by the network, and
t∗ = (t∗x , t∗y , t∗h , t∗w , t∗θ ) represents ground truth. λ is a balancing parameter that manages
the balance between class loss and regression loss. Only the regression loss is enabled if u
= 1 for the foreground and there is no regression for the background. The classification
loss function is defined as the cross-entropy loss between the ground-truth label u and the

10
predicted probability p:

Lcls (p, u) = −u · log(p) − (1 − u) · log(1 − p) (2)

Tuple t and t∗ are calculated like this:

tx = (x − xa )/wa , ty = (y − ya )/ha ,
tw = log(w/wa ), th = log(h/ha ), (3)
tθ = θ − θa

t∗x = (x∗ − xa )/wa , t∗y = (y ∗ − ya )/ha ,

t∗w = log(w∗ /wa ), t∗h = log(h∗ /ha ), (4)
t∗θ = θ∗ − θa
Where x, xa and x∗ denote predicted box, anchor and ground truth box, respectively; similar
for y, h, w and θ. The smooth-L1 loss is adopted for bounding box regression as follows:

u.smoothL1 (t∗i − ti )
X
Lreg (t, t∗) = (5)
i∈x,y,w,h,θ
(
0.5x2 if |x| <1
smoothL1 (x) = (6)
|x| − 0.5 otherwise

3.3.3 Box Head

The O-RPN network, as part of the architecture, generates 1000 proposal boxes and object-
less logits by default. These proposal boxes are then projected into the feature space using
an ROI pooling layer. The output of the ROI pooling layer is reshaped and fed into the fully
connected (FC) layers. The FC layers generate a RoI vector, which is then passed through
a predictor with two branches: the rotated bounding box regressor and the classifier. The
classifier, located in the classification layer of the model, predicts the object class, while the
regressor layer is responsible for regressing the bounding box values. This process enables
the model to accurately identify and classify objects within the input image, and also refine
the location of the proposed bounding boxes.

3.4 Evaluation Metrics

We used Mean Absolute Error, Error in Angle Prediction, fingerprint labeling accuracy,
and matching score to assess the effectiveness of the CRFSEG-v2 model on our contactless
fingerphoto dataset.

3.4.1 Mean Absolute Error

The Mean Absolute Error is a metric applied to evaluate the performance of fingerprint
segmentation models in accurately segmenting fingerprints within a specified geometric tol-
erance compared to human-annotated ground truth data.

11
 Negative error

B
Positive
A I error
C
E K
J
D Ground-truth
F
Predicted bounding box
bounding box

d
Predicte box
d in g
B boun
A -truth
C D Ground box
d in g
boun
F E

Figure 3: An example of calculating the positive and negative errors between the predicted
and ground truth bounding boxes involves using Euclidean distance. To determine the pixel
error forCenter
each side of the predicted bounding boxes, the distances between the endpoints
for Identification Technology Research – Spring 2021 Final Report
of a side and the corresponding endpoints of the annotated ground-truth bounding box
©CITeR NDA Terms Apply

© CITeR
are computed in pixels. For instance, when calculating the pixel error for the side AD,
perpendicular lines AC and DF are drawn with respect to the line AD. The perpendicular line
AC intersects the corresponding ground-truth line at point C, while the perpendicular line
DF intersects the extension of the corresponding ground-truth line at point F. Subsequently,
the Euclidean distances from point A to point C and from point D to point F are calculated.
The average of these two Euclidean distances represents the pixel error for the side AD. This
process is repeated for all four sides of the bounding box. Finally, Equation 7 is individually
applied to the pixel errors of each side, resulting in the calculation of the Mean Absolute
Error for that specific side.

Fingerprint segmentation models need to strike a balance between two factors: over-
segmentation and under-segmentation. Over-segmentation occurs when the predicted bound-
ing box is smaller than the ground truth fingerprint area, potentially capturing the ridge-
valley structure of adjacent fingerprints and introducing noise that may impact matching
performance. Under-segmentation, on the other hand, involves extending the predicted
bounding box beyond the actual fingerprint area, resulting in the loss of valuable finger-
print details and degrading matching performance. The MAE metric helps quantify the
extent of over-segmentation or under-segmentation produced by the model. To calculate the
Mean Absolute Error, we measure the distance in pixels between each side of the predicted
bounding box and the corresponding side of the annotated ground truth bounding box. A
successful segmentation refers to finding a bounding box around a fingerprint within a certain
geometric tolerance of the human-annotated ground truth bounding box.
Figure 3 illustrates more detail about calculating the MAE for a fingerprint. A detected
bounding box is considered to have a positive error if any of its sides encompass more data
than the corresponding side of the ground-truth bounding box. An example of a positive
error is shown on the right side of the rightmost fingerprint. A detected bounding box is

12
considered to have a negative error if any of its sides enclose less data than the corresponding
side of the ground-truth bounding box. An example of the negative error is shown on the
top side of the left-most fingerprint.
Finally, Equation 7 is applied independently to calculate the MAE for each side.
N
1X
MAE = |X errori | (7)
N i=0

N represents the total number of fingerprints within the dataset under evaluation. X signifies
the Euclidean distance error on any side of the bounding box, encompassing the top, bottom,
left, and right sides.

3.4.2 Error in Angle Prediction

Error in Angle (EAP) Prediction is employed to assess the ability of various fingerprint
segmentation systems to accurately predict the orientation of fingerprints. This involves
determining the deviation between the predicted angle of a fingerphoto and the ground
truth angle of the same fingerphoto, which is calculated using Equation 8.
N
1X
EAP = |θ − θ∗ | (8)
N i=0

Where N is the total number of fingerprints in the test dataset, θ is a ground truth angle
and θ∗ is a predicted angle by a fingerprint segmentation model.

3.4.3 Fingerprint classification accuracy

Hamming loss is employed to assess the performance of multi-class classifiers [19]. A pair
of sequences, one comprising the ground-truth label and the other containing the predicted
label, is evaluated using hamming loss, which quantifies the number of positions where the
corresponding symbols differ. Equation 9 is used to calculate the hamming loss [19].
N
1 X |Yi ∆Zi |
Hamming loss = (9)
N i=1 |L|

where N is total number of samples in dataset, and L is number of labels. Zi is the predicted
value for the i-th label of a given sample, and Yi is the corresponding ground true value. ∆
stands for the symmetric difference between two sets of predicted and ground truth values.
The accuracy of a multi-class classifier is related to Hamming loss [6, 11], which can be
computed using Equation 10.

Accuracy = 1 − Hamming loss (10)

3.4.4 Fingerprint Matching

Fingerprint matching is a crucial aspect of slap segmentation systems, where the performance
of the algorithms in correctly segmenting fingerprints within a specific tolerance is evaluated.

13
To assess fingerprint matching, we rely on the true accept rate (TAR) and false accept rate
(FAR). The TAR represents the percentage of instances in which a biometric recognition
system accurately verifies an authorized individual, calculated using Equation 11:
Correct accepted fingerprints
T AR = × 100% (11)
Total number of mated matching attempts
The false accept rate (FAR) measures the percentage of instances in which a biometric
recognition system mistakenly verifies an unauthorized user. It is computed using Equa-
tion 12.

Wrongly accepted fingerprints

F AR = × 100% (12)
Total number of non-mated matching attempts

3.5 Training
The CRFSEG-v2 model was developed using Detectron2, a framework created by Facebook
AI Research (FAIR) [22], which supports advanced deep learning-based object detection al-
gorithms. We utilized the Faster R-CNN algorithm and customized the Detectron2 code
to implement the oriented regional proposal network (ORPN), added new layers for han-
dling rotated bounding boxes, to meet our specific requirements for accurate slap fingerprint
segmentation.
In our experiments, we started with a pre-trained Faster R-CNN model trained on the
MS-COCO dataset, which has 81 output classes. However, since our task involved classifying
ten fingerprints from two hands, we adjusted the output layers to reduce the number of classes
from 81 to 10. The model was then fine-tuned using our unique slap image dataset. Training
followed an end-to-end strategy, where we calculated loss values by comparing the predicted
results against the ground truth. The training was conducted on a Linux-operated machine
equipped with a 20-core Intel(R) Xeon(R) E5-2690 v2 @ 3.00GHz CPU, 64 GB RAM, and
a NVIDIA GeForce 1080 Ti 12-GB GPU.
We performed a total of 40,000 training iterations for the fingerprint segmentation model.
The learning rates started at 10−4 and were decreased by a ratio of 0.1 at specific intervals
(4000, 8000, 12000, 18000, and 25000, 32000 iterations). The weight decay was set to 0.0005,
and the momentum was set to 0.7. Throughout the experiments, we employed multi-scale
training, eliminating the need for scaling the input before feeding it into the neural network.
The contactless fingerprint dataset, consisting of 23,650 finger photos, is divided using an
80:10:10 train/validate/test split ratio. A 10-fold cross-validation technique is employed to
construct and assess the model, and the outcomes are presented in the results section.

4 Results
This section presents a comprehensive analysis of our findings. We employed four distinct
metrics, namely Mean Absolute Error, Error in Angle Prediction, fingerprint labeling accu-
racy, and fingerprint matching accuracy, to assess the performance of our nobel CRFSEG-v2
segmentation model that handles contacless (contactbased) fingerprint images.

14
4.1 Mean Absolute Error of CRFSEG-v2 on our Contactless dataset
MAE measures the preciseness of bounding boxes around fingerprints generated by slap
segmentation algorithms. We used Equation 7 to calculate the MAE of our contactless
fingerphoto segmentation model on our novel dataset.
Table 2 presents the Mean Absolute Error and its corresponding standard deviation for
the segmentation model. The MAEs obtained for different sides of the predicted bounding
boxes are as follows: 26.09 for the left side, 27.33 for the right side, 20.23 for the top side, and
52.92 for the bottom side. It is worth noting that all these values are below the NIST-defined
tolerance threshold of 64 pixels. Furthermore, the achieved MAEs are comparable to those
obtained by the contact-based fingerprint segmentation system, indicating the effectiveness
of the proposed approach in achieving accurate segmentation results.

Table 2: The Mean Absolute Error (MAE) and its standard deviation were computed to
assess the performance of the contactless segmentation system on our contactless fingerphoto
dataset. The MAE was determined by averaging the absolute differences between each side
of the detected bounding box and the corresponding side of the ground-truth bounding
box, measured in pixels. A lower MAE value indicates better performance in accurately
segmenting the fingerphotos.
Dataset Side MAE(Std. dev.)
Left 26.09 (65.36)
Right 27.33 (64.29)
Contactless
Top 20.23 (52.97)
Bottom 52.92 (90.93)

Figure 4 showcases the histograms illustrating the Mean Absolute Error for all sides of
the bounding boxes generated by our contactless segmentation model. These histograms
are employed to showcase, analyze, and evaluate the MAE results. The generation process
involved subtracting the coordinate positions of the corresponding sides of the ground-truth
bounding boxes from the coordinate positions of the corresponding sides of the detected
bounding boxes obtained from the segmentation model. The histograms provide evidence of
improved performance in accurately segmenting contactless fingerphotos.

4.1.1 Error in Angle Prediction

The Error in Angle Prediction (EAP) refers to the discrepancy between the angles predicted
by the segmentation algorithms and the corresponding ground-truth angles of fingerphotos.
It directly reflects the precision of the bounding boxes detected by the segmentation model.
A smaller deviation among the predicted angles and the ground-truth angles signifies more
accurate segmentation outcomes. The calculation of EAP is performed using Equation 8.
For our contactless segmentation system, the resulting EAP is 5.92◦ , accompanied by a
standard deviation of 11.98◦ . The EAP values obtained from the segmentation model are
comparably lower and exhibit a similar trend to those of the contact-based slap fingerprint
segmentation, indicating superior performance.

15
(a) The Mean Absolute Error (MAE) in pixels (b) The Mean Absolute Error (MAE) in pixels for
for the left side of the fingerprints, as predicted the right side of the fingerprints, as predicted by
by the contactless segmentation system. the contactless segmentation system.

(c) The Mean Absolute Error (MAE) in pixels for (d) The Mean Absolute Error (MAE) in pixels for
the top side of the fingerprints, as predicted by the bottom side of the fingerprints, as predicted
the contactless segmentation system. by the contactless segmentation system.
Figure 4: The MAE histograms obtained from contactless fingerphoto segmentation algo-
rithm using Equation 7 are presented in this figure. Figure 4a, Figure 4b, Figure 4c, and
Figure 4d illustrate the MAE values for the left, right, top, and bottom sides of the bounding
box, respectively, as predicted by the contactless fingerphoto segmentation system.

16
 Figure 5 displays the histogram of the EAP generated by the contactless segmentation
model. The narrower spread of the histogram, as indicated by the smaller standard devia-
tion and area, demonstrates the superior performance of the segmentation model in angle
prediction.
The boxplot graph in Figure 6 displays the EAP values obtained from the contactless
segmentation model, illustrating the minimum, lower quartile, median, upper quartile, and
maximum values. The line extending across the box corresponds to the median value of the
EAP. Upon examining the boxplot graph, it becomes evident that the absolute median value
of the EAP generated by the contactless segmentation model is notably lower. This indicates
that the model demonstrates enhanced accuracy and precision in predicting the angle of
fingerphotos, even when they are excessively rotated. The boxplot graph illustrates the
constrained variability of the EAP values, further affirming the model’s ability to maintain
consistent performance regardless of the degree of rotation.

Figure 5: The histogram of the error in fingerphoto angle prediction by the contactless
segmentation model on the fingerphoto dataset. This error is calculated by subtracting the
angles predicted by the models from the ground-truth angles of the fingerphoto images. Low
standard deviation values indicate better results.

4.1.2 Label Prediction Accuracy of the contactless segmentation model

The fingerprint label prediction accuracy of the contactless segmentation model was cal-
culated using the Hamming Loss metric, which is widely used for evaluating multi-label
classifiers [9]. Our segmentation model demonstrated a remarkable label accuracy of 97.46%
on our contactless dataset, underscoring its superior accuracy in predicting fingerprint la-
bels. The model’s ability to accurately predict labels solely based on the characteristics of
the fingerprint images adds to its versatility and potential for broader application.

4.1.3 Fingerprint Matching

The main objective of a fingerprint segmentation algorithm is to improve the matching
performance. In our research, we specifically focus on accurately segmenting fingerphotos

17
Figure 6: Boxplots of the error in angle prediction (EAP) of the contactless segmentation
model on the contactless fingerphoto dataset. The boxplot statistical analysis of the mean
with ±10◦ (standard errors) for the EAP values of the model indicates that the algorithm
used in this model is invariant to the rotation of fingerphotos.

to achieve higher matching accuracy. After segmenting contactless fingerprints, we applied

various enhancement techniques. Utilizing the Verifinger 12.0 SDK, minutiae representations
are simultaneously retrieved from the contactless images [4]. We applied operations, such
as the Contrast Limited Adaptive Histogram Equalization (CLAHE) technique, to enhance
our contactless segmented fingerprint images [3]. To measure the matching accuracy, we
used the VeriFinger fingerprint matcher (version 10) which is compliant with the NIST
MINEX standards [20]. This allowed us to quantitatively evaluate the contributions of the
segmentation algorithm to the overall matching performance.
To calculate the matching accuracy, we generated two sets of segmented fingerprint im-
ages from the contactless fingerphoto dataset. One set of segmented fingerprints is generated
using the information of human-annotated (ground truth) bounding boxes and another set
is generated using the information of CRFSEG-v2 model (segmentation model) generated
bounding box information. We evaluated all possible genuine comparisons for each finger-
print in the segmented fingerprint set, while randomly selecting 100 non-mated fingerprints
to create an imposter distribution.
In Table 3, we present the matching performance for segmented fingerprints using both
the ground-truth bounding box information and the segmentation model. Our segmentation
model achieved a fingerprint-matching accuracy of 88.88%, whereas the ground-truth accu-
racy was 90.36%. However, It is crucial to remember that the fingerprint matching accuracy
for both the ground-truth data and the model’s output data falls short compared to the

18
contact-based fingerprint model, which achieves an accuracy of about 99%. Several factors
contribute to the lower matching results, which we thoroughly discuss in the discussion sec-
tion. Additionally, we provide suggestions for future research to overcome these challenges
and achieve better matching performance. In total, we conducted 6,208,760 comparisons
across all fingers in our experiment to estimate the matching accuracy.

Table 3: The True Positive Rate (TPR) at a False Positive Rate (FPR) of 0.001 is evaluated
for both ground-truth and contactless segmentation model segmented fingerprints within the
dataset. The results indicate CRFSEG-v2 performed close to the ground-truth level in terms
of fingerprint matching.
Model Accuracy
Ground-truth 90.36%
Contactless Model 88.88%

Figure 7 is the Receiver Operating Characteristics (ROC) curve that shows the matching
scores of our segmentation model CRFSEG-v2 along with ground truth on the contactless
dataset. The ROC curve is a graphical plot that represents the tradeoff between the true
positive rate (TPR) and the false positive rate (FPR) at various threshold values.

Figure 7: The Receiver Operating Characteristics (ROC) for the fingerprint matching per-
formance of Ground Truth, a newly developed fingerphoto segmentation model in the con-
tactless dataset.

19
5 Discussion
This study focuses on the development of a highly accurate slap segmentation system specif-
ically designed for contactless fingerphoto images. The system utilizes deep learning tech-
niques, particularly convolutional networks trained using an end-to-end approach to address
challenges such as rotation and noise commonly encountered in fingerphotos. One advan-
tage of our system is its ability to be fine-tuned using additional datasets, which enhances
its performance and generalization across a wide, diverse range of fingerphoto images. To
the best of our knowledge, commercial slap segmentation systems lack the flexibility to be
easily fine-tuned on diverse fingerphoto datasets.
The mean absolute errors (MAEs), error in angle prediction (EAP), and labeling accu-
racy achieved by our segmentation model are high and comparable to the accuracy levels
achieved by high-precision segmentation systems developed and tested using contact-based
slap images. However, we observed lower accuracy in the matching performance of our sys-
tem. It is crucial to remember that the matching performance heavily relies on the precise
segmentation of fingerphotos, accurate labeling of fingertips/fingerprints, and the quality of
the fingerphotos. Our segmentation model demonstrates the ability to effectively segment
fingerphotos even in cases where the image quality is poor, thereby achieving label accu-
racy that is nearly on par with human-level performance. However, when these segmented
images were subjected to different commercial matching software, we observed a decrease
in fingerprint-matching accuracy. Through manual examination, we determined that our
segmentation system accurately segments fingertips and assigns accurate labels to them. We
believe the reduced fingerprint matching accuracy can be attributed to the limitations of the
fingerprint matching software.

6 Conclusion
In this paper, we have examined the potential of contactless fingerprint authentication sys-
tems as a promising alternative to traditional contact-based methods. By leveraging deep
learning techniques, we have developed and evaluated a novel segmentation model for pre-
cise localization and extraction of contactless fingerprints. The results obtained from our
real-world dataset demonstrate the effectiveness and reliability of our novel segmentation and
extraction method. This work centers around the development of a segmentation model that
leverages deep learning techniques, specifically tailored for the segmentation of contactless
fingerprints.
Our novel CRFSEG-v2 model has achieved notable results, including an average Mean
Absolute Error (MAE) of 30 pixels, an Error in Angle Prediction (EAP) of 5.92 degrees, a
Labeling Accuracy of 97.46%, and a VeriFinger matching accuracy of 88.87%. Additionally,
we have curated an extensive in-house contactless dataset, comprising 23,650 finger photos.
While our research showcases promising results, there are still challenges to address in the
development of contactless fingerprint authentication systems. Future work should focus on
addressing issues like variability in image quality, occlusion, and lighting conditions to further
improve the robustness and generalization of the proposed system. With continued research
and advancements in deep learning techniques, we can expect even more sophisticated and

20
reliable contactless fingerprint authentication systems to emerge. These developments will
undoubtedly contribute to enhanced security and an improved user experience across a broad
range of applications, making fingerprint authentication an indispensable part of our digital
lives.
Here are some potential future steps that researchers can consider to address the matching
failure and hence improve performance:

• GAN-based approaches for enhancing low-quality images: Developing a generative

adversarial network (GAN) can help learn the mapping between low-quality fingerprint
images and their corresponding high-quality representations. By training the GAN on
a dataset containing pairs of low-quality and high-quality fingerprint images, it can
learn to enhance the low-quality images, thereby improving the matching accuracy.

• Feature extraction and matching algorithms: Robust feature extraction algorithms can
be employed to capture relevant fingerprint information, even in the presence of vari-
ations in image quality. These algorithms should be designed to handle distortions
caused by factors such as blurriness, missing parts, or poor lighting conditions. Sim-
ilarly, matching algorithms should be capable of accurately comparing and aligning
fingerprint features, even when dealing with imperfect images.

• Multiple capture and fusion: Instead of relying on a single fingerprint image, multiple
captures of the fingerprint can be obtained and then fused together. This approach
helps mitigate issues like missing parts or blurriness in individual images. Fusion tech-
niques such as averaging, weighted averaging, or feature-level fusion can be employed
to combine the information from multiple images and improve the overall matching
accuracy.

References
[1] Dragana Bartolić, Dragosav Mutavdžić, Jens Michael Carstensen, Slavica Stanković,
Milica Nikolić, Saša Krstović, and Ksenija Radotić. Fluorescence spectroscopy and
multispectral imaging for fingerprinting of aflatoxin-b1 contaminated (zea mays l.) seeds:
A preliminary study. Scientific Reports, 12(1):4849, 2022. [4]

[2] Samuel Cadd, Meez Islam, Peter Manson, and Stephen Bleay. Fingerprint composition
and aging: A literature review. Science & Justice, 55(4):219–238, 2015. [2]

[3] Sugandha Chakraverti, Pankaj Agarwal, Himansu Sekhar Pattanayak, Sanjay

Pratap Singh Chauhan, Ashish Kumar Chakraverti, and Manoj Kumar. De-noising
the image using dbst-lcm-clahe: A deep learning approach. Multimedia Tools and Ap-
plications, pages 1–26, 2023. [18]

[4] Steven A Grosz, Joshua J Engelsma, Eryun Liu, and Anil K Jain. C2cl: Contact
to contactless fingerprint matching. IEEE Transactions on Information Forensics and
Security, 17:196–210, 2021. [4, 5, 18]

21
 [5] Steven A Grosz and Anil K Jain. Afr-net: Attention-driven fingerprint recognition
network. IEEE Transactions on Biometrics, Behavior, and Identity Science, 2023. [2]

[6] Sooji Ha, Daniel J Marchetto, Sameer Dharur, and Omar I Asensio. Topic classification
of electric vehicle consumer experiences with transformer-based deep learning. Patterns,
2(2):100195, 2021. [13]

[7] Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. Deep residual learning
for image recognition. In 2016 IEEE Conference on Computer Vision and Pattern
Recognition (CVPR), pages 770–778, 2016. [9]

[8] Jun Jia, Guangtao Zhai, Jiahe Zhang, Zhongpai Gao, Zehao Zhu, Xiongkuo Min, Xi-
aokang Yang, and Guodong Guo. Embdn: An efficient multiclass barcode detection
network for complicated environments. IEEE Internet of Things Journal, 6(6):9919–
9933, 2019. [9]

[9] Sameh Khamis and Larry S. Davis. Walking and talking: A bilinear approach to multi-
label action recognition. In 2015 IEEE Conference on Computer Vision and Pattern
Recognition Workshops (CVPRW), pages 1–8, 2015. [17]

[10] Kenneth Ko. Users guide to export controlled distribution of nist biometric image
software (nbis-ec). 2007. [2]

[11] Oluwasanmi O Koyejo, Nagarajan Natarajan, Pradeep K Ravikumar, and Inderjit S

Dhillon. Consistent multilabel classification. Advances in Neural Information Processing
Systems, 28, 2015. [13]

[12] Aakarsh Malhotra, Anush Sankaran, Mayank Vatsa, and Richa Singh. On matching
finger-selfies using deep scattering networks. IEEE Transactions on Biometrics, Behav-
ior, and Identity Science, 2(4):350–362, 2020. [5]

[13] Davide Maltoni, Dario Maio, Anil K Jain, and Jianjiang Feng. Fingerprint sensing.
Handbook of Fingerprint Recognition, pages 63–114, 2022. [4]

[14] Emanuela Marasco and Anudeep Vurity. Fingerphoto presentation attack detection:
Generalization in smartphones. In 2021 IEEE International Conference on Big Data
(Big Data), pages 4518–4523, 2021. [2]

[15] MG Murshed, Keivan Bahmani, Stephanie Schuckers, and Faraz Hussain. Deep age-
invariant fingerprint segmentation system. arXiv preprint arXiv:2303.03341, 2023. [2,
4, 8]

[16] Tempestt J Neal and Damon L Woodard. Surveying biometric authentication for mobile
device security. Journal of Pattern Recognition Research, 1(74-110):4, 2016. [2]

[17] Martijn Oostdijk, Arnout van Velzen, Joost van Dijk, and Arnout Terpstra. State-of-
the-art in biometrics for multi-factor authentication in a federative context. Identity,
14:15, 2016. [2]

22
[18] Jannis Priesnitz, Rolf Huesmann, Christian Rathgeb, Nicolas Buchmann, and Christoph
Busch. Mobile contactless fingerprint recognition: implementation, performance and
usability aspects. Sensors, 22(3):792, 2022. [2]

[19] Grigorios Tsoumakas and Ioannis Manousos Katakis. Multi-label classification: An

overview. Int. J. Data Warehous. Min., 3:1–13, 2007. [13]

[20] CI Watson, G Fiumara, E Tabassi, S Cheng, P Flanagan, and W Salamon. Fingerprint

vendor technology evaluation, nist interagency/internal report 8034: 2015, 2014. [18]

[21] Peter Wild, Franz Daubner, Harald Penz, and Gustavo Fernández Domínguez. Com-
parative test of smartphone finger photo vs. touch-based cross-sensor fingerprint recog-
nition. 2019 7th International Workshop on Biometrics and Forensics (IWBF), pages
1–6, 2019. [4]

[22] Yuxin Wu, Alexander Kirillov, Francisco Massa, Wan-Yen Lo, and Ross Girshick. De-
tectron2. 2019. 2019. [14]

7 Acknowledgements
This material is based upon work supported by the Center for Identification Technology
Research and the National Science Foundation under Grant Number 1650503.

23