1. 1
DESIGN OF A HAND GEOMETRY BASED BIOMETRIC
SYSTEM
A seminar report
Submitted in the partial fulfillment of the requirements for the award of
Degree of
Master of Engineering
In
Electronics Instrumentation & Control Engineering
Submitted By:
******
Roll No.********
Under the supervision of:
**********
DEPARTMENT OF ELECTRICAL AND INSTRUMENTATION
ENGINEERING
*************
2. 2
ABSTRACT
Biometrics which can be used for identification of individuals based on their
physical or behavioral characteristics has gained importance in today’s society
where information security is essential. Hand geometry based biometric systems
are gaining acceptance in low to medium security applications. Hand geometry
based identification systems utilize the geometric features of the hand like length
and width of the fingers of the hand. The proposed system is a verification system
which utilizes these hand geometry features for user authentication. The system
accepts a grayscale handprint from which it extracts the finger lengths and finger
widths.
3. 3
ACKNOWLEDGEMENT
The real spirit of achieving a goal is through the way of excellence and austerous
discipline. I would have never succeeded in completing my task without the
cooperation, encouragement and help provided to me by various personalities.
First of all, I render my gratitude to the ALMIGHTY who bestowed self-
confidence, ability and strength in me to complete this work. Without his grace this
would never come to be today’s reality.
With deep sense of gratitude I express my sincere thanks to my esteemed and
worthy supervisor Dr. Sunil Kumar Gupta for his invaluable guidance. It would
have never been possible to complete this work without his continuous support and
encouragement. Inspite of his very busy schedule he was always approachable and
available to attend to my problems, discuss the solutions, and give the appropriate
advice. Doing research under his supervision was a very enlightening and
enjoyable experience. I also wish to thank all the faculty members of the
Department of Electrical and Instrumentation Engineering for the invaluable
knowledge they have imparted on me and for teaching the principles in most
exciting and enjoyable way. I also extend my thanks to the technical staff of the
department for maintaining an excellent working facility. I would like to thank my
parents for supporting and taking me to this stage in life; it was their blessings
which always gave me courage to face all challenges and made my path easier.
**********
Roll No.********
4. 4
TABLE OF CONTENTS
CH. NO. TITLE PAGE NO.
Abstract 2
Acknowledgement 3
Table of contents 4
List of figures 5
1. Introduction 6-10
2. Literature review 11-16
3. Comparison of various biometric technologies 17-19
4. Proposed Work 20-25
- Methodology
Image Acquisition
Image Preprocessing
Hand Feature Extraction
Matching
Decision
5. References 26-28
5. 5
LIST OF FIGURES
S.No. Figure Page no.
1. Typical architecture of hand geometry biometrics 8
2. Diagram of a general biometric system 10
3. Equal Error Rate 15
4. Input image 22
5. Image after Binarization 22
6. Binarized image 24
7. Feature extraction 24
8. Template matching against templates in a database 25
7. 7
1. Introduction
Biometrics, by definition, is a word whose origins related with the ancient greek
language. "Bios" which means life and "metron" which means measure. For that
reason biometrics is the emerging technology for distinguishing individuals based
upon recognition of one or more traits. Physical and behavioral characteristics are
acquired by Biometric Systems. Biometric technologies are becoming the
foundation of an extensive array of highly secure identification and personal
verification solutions.
Typical architecture of all biometric systems consists of two phases:
• Enrollment,
• Recognition.
In the phase of enrollment, several images of hand are taken from the users. The
images, called templates, are preprocessed to enter feature extraction, where a set
of measurement is performed.
Final model depends on the method used for recognition. Models for each of the
users is then stored in the database. In the phase of recognition, a single picture is
taken, preprocessed, and features are obtained. In the proposed system, the process
of verification is used, where the input template is compared only with the model
of claimed person. The feature vector is compared with features from the model
previously stored in the database. The result is the person is either authorized or
not authorized.
9. 9
Traditional hand geometry-based systems use low-resolution cameras or scanners
to capture users’ hand images with the help of peggies or by forcing them to touch
a screen. Those systems measure a hand shape to extract its features, like lengths
and widths of fingers, and hand contour, for recognition. Unfortunately, traditional
techniques face unsolved problems: low discriminability due to low-resolution
hand images and bad user acceptability because users worry about hygienic issues
when they have to touch screens.
Because of the increased hygiene concern in biometric systems and the difficulty
in recognizing fingerprints of manual laborers and elderly people, hand geometry
has been currently employed in many systems for personal verification mostly as a
complement to finger-print authentication.
To improve discriminability and user acceptability, our new hand recognition
systems have to acquire high-resolution hand images without peggy constraints
and also contacts. However, those new hand images rise new challenges, like hand
texture, motion and shadow, to extract hand shapes, and measure hand features.
Our project aims at those new challenges and gives a machine learning solution.
A biometric system could have either or both of the two features, Identification and
Verification. In the process of identification the individual presents the required
biometric characteristic and the biometric system associates an identity to that
Individual. In the case of recognition or verification, however, the person presents
requires both the biometric characteristic and an identity. The system then verifies
whether that identity is associated with that person’s biometric characteristic or
not. The proposed work aims to perform verification.
12. 12
2.1LITERATURE REVIEW
Hand geometry is a biometric that identifies users by the shape of their hands.
Hand geometry readers measure a user's hand along many dimensions and compare
those measurements to measurements stored in a file.
Viable hand geometry devices have been manufactured since the early 1980s,
making hand geometry the first biometric to find widespread computerized use. It
remains popular; common applications include access control and time-and-
attendance operations.
Since hand geometry is not thought to be as unique as fingerprints or irises,
fingerprinting and iris recognition remain the preferred technology for high-
security applications. Hand geometry is very reliable when combined with other
forms of identification, such as identification cards or personal identification
numbers. In large populations, hand geometry is not suitable for so-called one-to-
many applications, in which a user is identified from his biometric without any
other identification.
Nowadays it is easy to find biometric devices providing physical access to places
or logical access to computer data in several places from large companies to small
gyms. Hand geometry is a kind of biometric measure that is not as diffused in the
market as others, nevertheless in 2004 it took 11% of the entire market share for
biometric technologies (according to the International Biometric Group).
Differently from most of the other systems [1], this work provides a new approach
to the way hand geometry features are extracted. Data is read and processed
independently of the position of the user hand. This is done by analyzing the
curvature profile of the hand contour, making the feature extraction process
rotation and translation invariant. Hand Geometry based human verification
technique which is efficient, simple, fast, easy to handle and cost effective
compared to other verification techniques.
Hand Geometry is a biometric key with medium level of individualization.
Experiments show that the physical dimensions of a human hand contain
information that is capable to verify the identity of an individual. There are several
features that can be extracted and used as key such as finger width and length,
13. 13
overall size of the hand, hand contour among others (a deep study about hand
geometry features can be seen in [2]).
Hand-based authentication schemes in the literature are mostly based on
geometrical features.
For example, Sanchez-Reillo et al. [3] measure finger widths at different latitudes,
finger and palm heights, finger deviations and the angles of the inter-finger valleys
with the horizontal. The twenty-five selected features are modeled with Gaussian
mixture models specific to each individual.
Öden, Erçil and Büke [4] have used fourth degree implicit polynomial
representation of the extracted finger shapes in addition to such geometric features
as finger widths at various positions and the palm size. The resulting sixteen
features are compared using the Mahalanobis distance.
Jain, Ross and Pankanti [5] have used a peg-based imaging scheme and obtained
sixteen features, which include length and width of the fingers, aspect ratio of the
palm to fingers, and thickness of the hand. The prototype system they developed
was tested in a verification experiment for web access over for a group of 10
people.
Bulatov et al. [6] extract geometric features similar to [3,4,5] and compare two
classifiers.
The method of Jain and Duta [7] is somewhat similar to that they compare the
contour shape difference via the mean square error, and it involves fingers
alignment.
Lay [8] introduced a technique where the hand is illuminated with a parallel
grating that serves both to segment the background and enables the user to register
his hand with one the stored contours. The geometric features of the hand shape are
captured by the quadtree code.
Finally let’s note that there exist a number of patents on hand information-based
personnel identification, based on either geometrical features or on hand profile
[9].
14. 14
2.2. Advantages
• Acquisition convenience and good verification performance
• Suitable for medium and low security applications.
• Ease of integration.
• Currently being used for functions such as access control, employee time
recording and point of sale applications.
• Reasonably high acceptance among users and it is opt-in.
• Works in challenging environments.
• Low template size, which reduces storage needs.
2.3. Disadvantages
• Large size of hand geometry device.
• Limiting the applications of hand geometry system to verification task only.
• Single hand use only.
• It is not highly unique.
• Weather, temperature and medical conditions such as pregnancy or certain
medication can affect hand size.
• Hand size and geometry changes over time, especially in the very young and very
old.
15. 15
2.4. United States Government Evaluations
The US government has sponsored two evaluation of hand geometry technology.
The 1996 Evaluation of the INSPASS Hand Geometry Data determined the effect
of a threshold on system operation, established false accept and false reject rates as
a function of the threshold, and presented and estimate of the Receiver Operating
Characteristics (ROC) curve for the INSPASS system [10]. The evaluators noted
that an estimate was the best that could be achieved with the available data. A 1991
Performance Evaluation of Biometric Identification Devices evaluated the relative
performance of multiple biometric devices, including hand geometry [11].
The performance of a biometric system is measured in certain standard terms.
These are false acceptance rate (FAR), false rejection rate (FRR) and equal error
rate (EER) also called crossover error rate (CER).
FAR is the ratio of the number of unauthorized (unregistered) users accepted by
the biometric system to the total of identification attempts made.FRR is the ratio of
the number of number of authorized users rejected by the biometric system to the
total number of attempts made. Equal error rate is a point where FRR and FAR are
same.
Figure3.Equal Error Rate
16. 16
False acceptance poses a much more serious problem than false rejection. It is
therefore desired that the biometric system keep the FAR to the minimal possible
limit. This can be achieved by setting a high threshold so that only very near
matches are recognized and all else are rejected. The higher the security
requirement from the system the higher the threshold required to maintain it.
However FRR also depends upon the threshold. As the threshold increases the
FRR increases proportionally with it. This is because due to a high threshold
matches which are correct but below the threshold due to noise or other factors will
not be recognized. It is therefore desired that a balance is maintained. Usually this
balance point is the ERR where the FRR and the FAR are equal. However the
security requirements from the system are the primary concern while deciding the
threshold value and either of the FAR or FRR might be sacrificed for the other. In
case of a very high security system the threshold may be raised while for a system
where false rejects are of more concern the threshold might be lowered.
18. 18
3.1 Comparison of Various Biometric Technologies
The choice of a particular human characteristic to be used as a biometric trait
depends on the following criteria:
Uniqueness is how well the biometric separates individually from another.
Permanence measures how well a biometric resists aging.
Collectability ease of acquisition for measurement.
Performance accuracy, speed, and robustness of technology used.
Acceptability degree of approval of a technology.
Circumvention ease of use of a substitute.
The following table shows a comparison of existing biometric systems in
terms of those parameters [12]. A low ranking indicates poor performance in the
evaluation criterion whereas a high ranking indicates a very good performance.
Table1. Comparison between Different Biometrics
Technology
characteristic
Fingerprint Iris Facial Hand
How it works Capture
And compares
Fingertip
patterns
Capture
And compares
iris patterns
Capture
And compares
facial patterns
Measures
And compares
dimensions of
hand and
fingers
Cost of device Low High Moderate Moderate
Enrollment
Time
About 3
minutes,
30 seconds
About 2
minutes, 15
seconds
About 3
minutes
About 1minute
Transaction
time
9-19 seconds 12 seconds 10 seconds 6-10 seconds
19. 19
False non
match rate
.2%-36% 1.9%-6% 3.3%-70% 0%-5%
False match
rate
0%-8% <1% .3%-5% 0%-2.1%
Variability
with ages
Stable Stable Affected by
aging
Stable
Commercial
Availability
since
1970s 1997s 1990s 1970s
As can be seen in this Table, each and every individual technology has limitation
either in universality, uniqueness, permanence, collectability, or performance,
acceptability, circumvention. Due to these limitations, no single biometric can
provide a desired performance and the usage of multimodal biometric traits sounds
promising.. Exploiting information from multiple biometric sources or features
improves the performance and also robustness of person authentication [13].One of
most widely reported multimodal biometric authentication is combination of
speech and signature features. Research shows that they result in good
performance, but limited applications. Perhaps they didn’t collect the data from
practical environment. So, that’s still far from public applicability.
21. 21
4. METHODOLOGY
Hand geometry recognition is based on the extraction of a hand pattern that
incorporates parameters such as finger length, width, thickness, curvatures, or
relative location. Hand geometry refers to the geometric structure of a hand, which
includes lengths of fingers, widths at various points on the finger, diameter of the
palm, thickness of the palm, etc. [14]. These features are not as discriminating as
other biometric characteristics (such as fingerprints), however they can easily be
used for verification purpose.
The algorithm to extract the feature involves the following steps:
Image Acquisition
Image Preprocessing
Hand Feature Extraction
Matching
Decision
4.1 Image Acquisition
Image acquisition is the first step in a hand geometry biometrics system. The
image acquisition involves capturing and storing digital images from vision
sensors like color digital cameras, monochrome and color CCD cameras, video
cameras, scanners, etc.The image acquisition system comprises of a light source ,a
digital camera/scanner. The input image is a color/grayscale image of a hand. In
the proposed system images are acquired through a digital camera. It is necessary
that the fingers are separated from each other. However it is not required to stretch
the fingers to far apart as possible. The hand should be placed in a relaxed state
with fingers separated from each other. Since features such as length and width
which are dependent on the image size and resolution are being used, it is critical
that to have uniform size of images.
There are various formats stored for the images such as .jpeg,.tiff,.png,.gif and
.bmp. The captured images are stored in one of the following formats on the
computer for possible image processing. The input image, shown in Figure3 is
stored in .png format.
22. 22
4.2 Image Preprocessing
The next stage is image preprocessing module. Image preprocessing relates to the
preparation of an image for later analysis and use. Images captured by a camera or
a similar technique are not necessarily in a form that can be used by image analysis
routines. Some may need improvement to reduce noise; other may need to be
simplified, enhanced, altered, segmented, filtered, etc.The role of the processing
module is to prepare the image for feature extraction.
Image preprocessing module consists of following operations:
(i) Gray scale image
(ii) Noise removal
(iii) Edge detection
(i) Grayscale image
The first step in the preprocessing block is to transform the color image into a
grayscale image. A red, green and blue (RGB) value of each pixel is extracted.
Since a monochromatic image is required for the proposed system a threshold is
determined. All pixels with RGB values above the threshold are considered white
pixels and all pixels below the threshold are considered black pixels. Initially the
threshold is set very low, very close to the RGB value of a black pixel in the
image. This produces an image with a completely white palm on a black
background as shown in Figure 4. Features such as finger lengths perimeter and
area of the palm can be more easily extracted from this image. However setting the
threshold very low results in a lot of noise in the image. A good threshold is
determined and then noise removal algorithms are applied to the image.
Figure 4: input image Figure 5: Image after Binarization
23. 23
(ii) Noise Removal
Ideally the scanned input image should contain no noise. However due to dust and
dirt both on the palm and on the scanner bed, even in minute quantities may
produce differences between the actual image scanned and the palm print. These
variations may also be the result of a host of other factors including the settings of
the scanner, the lighting effects, humidity in the atmosphere etc. These variations
unless removed adversely affect the performance of the system. The larger the
degree of variations or noise the less accurate the system. So before extracting
features from the image, noise is reduced as much as possible. However most noise
removal algorithms also affect the actual features so a balanced approach is taken
such that the features are undamaged after noise removal. Background lightning
effects and the noise make fake pixels in the image. MATLAB function imfilter is
used to remove these pixels and to justify edges of the hand in the next step. The
function provides filtering of multidimensional images. The imfilter function
computes the value of each output pixel using double-precision, floating-point
arithmetic. Input image pixel values outside the bounds of the image are assumed
to equal to the nearest array border value. Hand boundary is easily located
afterwards.
(iii) Edge Detection
In order to extract geometric features of the hand it is required that the image
contains only edges. Edge detection is the process of localizing pixel intensity
transitions. The edge detection has been used by object recognition, target tracking,
segmentation, etc.An edge is a collection of connected high frequency points in an
image.Visually,an edge is a region in an image where there is a sharp change in
intensity of an image. Detecting edges of an image represents significantly
reduction in the amount of data and filters out useless information, while
preserving the important structural properties in an image.
Therefore, the edge detection is one of the most important parts of image
preprocessing. There mainly exist several edge detection methods like Sobel,
Prewitt, Robets, and Canny.
24. 24
4.3 Hand Feature Extraction
There are several features that can be extracted from the geometry of the hand.
Each finger has three major lines running perpendicular to the length of the finger.
The first feature that can be extracted is the length of a finger which is defined as
the distance between the tip of the finger and bottommost line on the finger. The
second major feature is the width of the finger. One or more measurements can be
taken for the width at varying points along the finger. The length of the lines on the
finger can also be used as the measure of finger width. Since the fingers may not
have uniform width usually two or more measurements are taken for each finger
along different points.
Figure 6: Binarized image Figure7: Feature Extraction
4.4 Matching
The matching stage provides the means to determine the identity of a user. When a
user attempts recognition in a biometric system, the user’s generated features
template will be compared against the templates stored in the database.
In one-to-one verification, this comparison is done only against the claimed
identity’s template, whereas in a one-to-many identification it is done against the
entire database. Since the case of verification is just a subset of the identification
25. 25
case, only the later is described and reported in this work – verification will
typically yield better results.
The matching stage is based on a classification algorithm that generates a distance
score for each template comparison using a feature vectors’ similarity measure.
The score with the lowest distance value indicates the best match. Unnecessary
template matching comparisons are avoided by also taking into account if the
templates being compared both belong to the right or left hand, information which
is obtained from the pre-processing stage.
4.5 Decision
After running the matching algorithm, a recognition decision is made whether to
accept or reject the best match found. If the distance score exceeds a predefined
threshold, the recognition attempt is considered as an impostor access, otherwise
the recognition attempt is considered a client access and the system assumes the
user has been correctly identified. The classification procedure is illustrated in
Figure 9.
Figure 7: Template matching against templates in a database.
27. 27
[1] R. Sanchez-Reillo, C. Sanchez-Avila, A. Gonzalez-Marcos,“Biometric
Identification through Hand Geometry Measurements”, IEEE Trans. On Pattern
Analysis and Machine Intelligence v22 n10, pp. 1168-1171, 2000.
.
[2] S; Travieso C.M.; Alonso, J.B.; Ferrer M.A.; "Automatic biometric
identification system by hand geometry", IEEE 37th Annual 2003 International
Carnahan Conference on 14-16, 281 – 284, Oct. 2003.
[3] R. Sanches-Reillo, C. Sanchez-Avila, and A. Gonzalez-Marcos, “Biometric
Identification through Hand Geometry Measurements,” IEEE Transactions of
Pattern Analysis and Machine Intelligence, Vol. 22, No. 10, October 2000.
[4] C. Öden, A. Erçil and B. Büke, "Combining implicit polynomials and
geometric features for hand recognition", Pattern Recognition Letters, 24,
2145-2152, 2003. .
[5] A.K. Jain, A. Ross and S. Pankanti, "A prototype hand geometry based
verification system", Proc. of 2nd Int. Conference on Audio- and Video-
Based Biometric Person Authentication, pp.: 166-171, March 1999.
[6] Y. Bulatov, S. Jambawalikar, P. Kumar and S. Sethia, "Hand recognition
using geometric classifiers", DIMACS Workshop on Computational
Geometry, Rutgers University, Piscataway, NJ, November 14-15, 2002.
[7] A.K. Jain and N. Duta, "Deformable matching of hand shapes for
verification, Proc. of Int. Conf. on Image Processing, October 1999.
[8] Y. L. Lay, “Hand shape recognition,” Optics and Laser Technology, 32(1),
1–5, Feb. 2000.
[9] R.L. Zunkel, “Hand Geometry Based Verification”, pp. 87-101, in
Biometrics, Eds. A. Jain, R. Bolle, S. Pankanti, Kluwer Academic
Publishers, 1999.
[10] James Wayman, ed., “National Biometric Test Center Collected Works,”
San Jose State University, August 2000.
http://www.engr.sjsu.edu/biometrics/nbtccw.pdf.
28. 28
[11] James Holmes, Larry Wright, and Russell Maxwell,” A Performance
Evaluation of Biometric Identification Devices,” Sandia National Laboratories,
June1991.http://infoserve.sandia.gov/cgi-bin/techlib/access-control.pl/1991/
910276.pdf.
[12] Comparisons of Various Biometric Technologies,
www.biometricvision.com
[13] M. N. Eshwarappa and M. V. Latte, Bimodal Biometric Person
Authentication System Using Speech and Signature Features, International Journal
of Biometrics and Bioinformatics, (IJBB), Volume (4): Issue (4)
[14] R. S. Chora´s, M. Chora´s, ”Multimodal Hand-Palm Biometrics,” in
Adaptive and Natural Computing Algorithms, Springer Verlag LNCS 4432, pp.
407 - 414.