BIOMETRICS
`Personal Identification
`
`in Networked Society
`
`edited by
`
`Anil Jain
`
`Ruud Bolle
`
`Sharath Pankanti
`
`
`
`APPLE 1112
`
`

`

`BIOMETRIC S
`
`Personal Identification
`in Networked Society
`
`

`

`—“
`
`
`
`THE KLUWER INTERNATIONAL SERIES IN ENGINEERING
`
`AND COMPUTER SCIENCE
`
`

`

`iBIOMETRICS/
`Personal Identification
`in Networked Society
`
`edited by
`
`Anil K. Jain
`
`Michigan State University
`ELansing, Michigan
`
`and
`
`Ruud Bolle and Sharath Pankanti
`
`IBM TJ .Watson Research Center
`Yorktown Heights, New York
`
`Kluwer Academic Publishers
`
`Bos ton/Dordrech t/London
`
`é
`
`

`

`
`
`
`
`Distributors for North, Central and South America:
`Kluwer Academic Publishers
`
`101 Philip Drive
`Assinippi Park
`Norwell, Massachusetts 02061 USA
`Tel: 781-871—6600
`Fax: 781-871-6528
`
`E-mail: kluwer@wkap.com
`
`Distributors for all other countries:
`Kluwer Academic Publishers Group
`Distribution Centre
`Post Office Box 322
`3300 AH Dordrecht, THE NETHERLANDS
`Tel: 31 78 6392 392
`Fax: 31 78 6546 474
`
`E-mail: orderdept@wkap.nl
`
`r.
`
`.
`'
`n r:
`E K ”i (:5) .j
`v.1
`
`‘
`
`P ”4.2
`__*"
`\*
`__
`
`'
`
`‘
`
`~
`
`‘
`
`~
`
`efi
`
`Electronic Services: http://www.wkap.nl
`
`Library of Congress Cataloging-in-Publication Data
`
`identification in networked society / edited by
`: personal
`Biometrics
`Anil K. Jain and Ruud Belle and Sharath Pankanti.
`
`(The Kluwer international series in engineering and
`——
`cm.
`p.
`computer science ; SECS 479)
`Includes bibliographical references and index.
`ISBN 0-7923-3345-1
`
`3.
`2. Biometry.
`1. Pattern recognition systems.
`I. Jain, Anil K.,
`—Automation.
`4. Computer vision.
`1]. Belle, Ruud.
`III. Pankanti, Sharath.
`IV. Series.
`TK7882P3B36
`1998
`006.4--dc21
`
`Identification—
`1948—
`
`98—31272
`CIP
`
`
`Capyright © 1999 by Kluwer Academic Publishers. Second Printing 1999.
`
`All rights reserved. No part of this publication may be reproduced, stored in a
`retrieval system or transmitted in any form or by any means, mechanical, photo-
`copying, recording, or otherwise, without the prior written permission of the
`publisher, Kluwer Academic Publishers, 101 Philip Drive, Assinippi Park, Norwell,
`Massachusetts 02061
`
`Printed on acid-flee paper.
`
`Printed in the United States of America
`
`
`
`M.|.T. LIBRARIES
`
`11f "—‘mf
`F E B 1 8 2000
`i\
`_._....__...i
`RECEIVEDHum
`
`

`

`Contents
`
`Foreword
`
`Preface
`
`1.
`
`Introduction to biometrics
`
`A- K. Jain, R. Bolle, and S. Pankanti
`
`2.
`
`3.
`
`Fingerprint verification
`L. O’Gorman
`
`Face recognition
`J. Weng and D. L. Swets
`
`4. Hand geometry based verification
`R. Zunkel
`
`5. Recognizing persons by their Iris patterns
`J. Daugman
`
`6. Retina identification
`'R. Hill
`
`7. Automatic on—line signature verification
`V. Nalwa
`
`8.
`
`9.
`
`Speaker recognition
`J. Campbell
`
`Infrared identification of faces and body parts
`F. J. Prokoski and R. Rjedel
`
`-’\
`
`10. Keystroke dynamics based authentication
`M. S. Obaidat and B. Sadoun
`
`11. Automatic gait recognition
`M. S. Nixon, J. N. Carter, D. Cunado, PS. Huang, and S. V. Stevenage
`
`12. Objective odour measurements
`K. C. Persaud, D—H. Lee, and H-G. Byun
`
`13. Ear biometrics
`
`M- Burge and W. Burger
`
`14. DNA based identification
`N. Rudin, K. Inman, G. Stolovizky, and I. Rigoutsos
`
`15. Large scale systems
`R. Germain
`
`l6. Multimodal biometrics
`
`L. Hong and A. Jain
`
`17. Technical testing and evaluation of biometric identification devices
`J. L. Wayman
`
`vii
`
`43
`
`65
`
`87
`
`103
`
`123
`
`143
`
`165
`
`191
`
`213
`
`231
`
`251
`
`273
`
`287
`
`311
`
`327
`
`345
`
`

`

`vi
`
`Biometrics
`
`18. Smancard based authentication
`
`N. K. Ratha and R. Bolle
`
`19. Biometrics: Identifying law and policy concerns
`J. Woodward
`
`Index
`
`369
`
`385
`
`407
`
`

`

`Foreword
`
`The need to authenticate ourselves to machines is ever increasing in today’s
`networked society and is necessary to close the air gap between man and machine to
`secure our transactions and networks. Only biometrics (automatically recognizing a
`person using distinguishing traits) can recognize you as you.
`This first book on biometrics advances the science of biometrics by laying a
`- foundation and theoretical framework in contributed chapters by leading experts from
`industry and academia. Biometric technology has advanced tremendously over the last
`few years and has moved from research labs and Hollywood to real—world
`applications. Like any technology with commercial applications, it has been difficult,
`until now, to assess the state of the art in biometrics in the open literature.
`As Mark Twain said, “First get your facts; then you can distort them at your
`leisure.” We are moving away from the Twain era of evaluation to the science of
`evaluation with a
`chapter on scientifically based performance
`evaluation.
`Understanding the performance of biometric systems in various real-world situations
`is key to their-application.
`Most of today’s major biometric technologies are represented here by chapters on
`face,
`fingerprint, and speaker recognition, among others.
`In keeping with the
`networked society theme of this book, included are chapters on systems, networking,
`related technologies, and privacy issues. After the benefits of biometrics exceed their
`cost and their performance is understood, social, legal, and ethical issues are crucial to
`society’s accepting the ubiquitous application of biometrics. This book encompasses
`all these aspects of biometrics, which are introduced by the editors.
`
`Joseph P. Campbell, Jr.
`Chair, Biometric Consortium, 1994-1998
`http://wwwbiomen'icorgf
`
`

`

`Preface
`
`0
`
`Determining the identity of a person is becoming critical in our vastly interconnected
`information society. As increasing number of biometrics—based identification systems
`are being deployed for many civilian and forensic applications, biometrics and its
`applications have evoked considerable interest. The current state of affairs is that the
`technical and technological literature about the overall state-of-the-art in biometrics is
`dispersed across a wide spectrum of books, journals, and conference proceedings. As
`biometrics emerges as a multi-billion dollar industry, there is a growing need for a
`comprehensive, consolidated,
`fair, and accessible overview of the biometrics
`technology and its implications to society from well-reputed information sources.
`This edited book is an attempt
`to disseminate the technological aspects and
`implications of biometrics. In particular, this book addresses the following needs.
`0
`Survey the biometrics methods in commercial use and in research stage.
`«- Assess the capabilities and limitations of different biometrics.
`I
`Understand the general principles of design of biometric systems and the
`underlying trade-offs.
`0 Understand the issues underlying the design of biometric systems.
`0
`Identify issues in the realistic evaluation of biometrics—based systems.
`0
`To recognize personal privacy and security implications of biometrics-based
`identification technology.
`To nurture synergies of biometric technology with the other existing and
`emerging technologies.
`The book is organized as follows: Chapter 1 is a brief overview of the biometric
`technology and the research issues underlying the biometrics-based identification
`applications. A number of biometrics-based technologies are commercially available
`today and many more are being develoPed in the educational and commercial research
`laboratories world wide. Currently,
`there are mainly eight different biometrics
`including face, fingerprint, hand geometry, iris, retinal pattern, signature, voice—print,
`and thermograms have actually been deployed for identification.
`In each of the next
`eight chapters
`(Chapters 2—9),
`the leading experts and pioneers of biometric
`technology describe a particalar biometric, its characteristics, the specific problems
`underlying the design of an identificationlauthentication system based on that
`biometric, performance evaluation of the existing systems and open issues which need
`to be addressed. The next five chapters (Chapters 10-14) describe biometrics which
`are not yet commercially available but which are under active research for on-line
`identification: keystroke dynamics, dait, odor, ear, and DNA.
`A number of emerging civilian applications involve a very large number of
`identities (e.g., several million) and at the same time have demanding performance
`requirements (e.g., scalability, speed, accuracy). Chapter 15 addresses the research
`issues underlying design of a large identification and authentication system. To
`accomplish and engineer the design of highly reliable and accurate biometrics-based
`identification systems,
`it may often be
`necessary to effectively integrate
`discriminatory information contained in several different biometrics.
`These
`integration issues are dealt with in Chapter 16. A large cross-section of the population
`interested in biometrics is overwhelmed by the quickly growing pace of the
`
`

`

`

`

`1 INTRODUCTION TO BIOMETRICS
`
`Anil Jain
`Michigan State University
`East Lansing, MI
`jain@cse.msu.edu
`
`..
`
`Ruud Bolle and Sharath Pankanti
`IBM T. J. Watson Research Center
`Yorktown Heights, NY
`{bolle,sharat}@us.ibm.com
`
`Abstract Biometrics deals with identification of individuals
`based on their biological or behavioral characteristics. Biometrics
`has lately been receiving attention in popular media. it is widely
`believed that biometrics will become a significant component of the
`identification technology as (t) the prices of biometrics sensors
`continue to fall,
`(it)
`the underb’ing technology becames more
`mature. and (iii) the public becomes aware of the strengths and
`limitations of biometrics. This chapter provides an overview of the
`biometrics technology and its
`applications and introduces the
`research issues underlying the biometrics.
`
`access
`venfication,
`identification,
`Keywords: Biometrics,
`control,
`authentication,
`secun'ty,
`research issues,
`evaluation,
`privacy.
`
`1 .
`
`Introduction
`
`identification. The
`is called personal
`Associating an identity with an individual
`problem of resolving the identity of a person can be categorized into two
`fimdamentally distinct types of problems with different inherent complexities: (i)
`verification
`and (ii)
`recognition (more popularly known as
`identificationl).
`Verification (authentication) refers to the problem of confirming or denying a person's
`
`I The term identification is used in this book either to refer to the general problem of identifying
`individuals (identification and authentication) or to refer to the specific problem of identifying an
`individual from a database which involves one to many search. We rely on the context to disambiguate the
`reference.
`
`

`

`2
`
`Jain et a1.
`
`Identification (Who am I?) refers to the
`claimed identity (Am I who I claim I am?)
`problem of establishing a subject's identity - either from a set of already known
`identities (closed identification problem) or otherwise (open identification problem).
`The term positive personalr identification typically refers (in both verification as well
`as identification context) to identification of a person with high certainty.
`Human race has come a long way since its inception in small tribal primitive
`societies where every person in the community knew every other person. In today's
`complex,
`geographically mobile,
`increasingly
`electronically
`inter-connected
`information society, accurate identification is becoming very important and the
`problem of identifying a person is becoming ever increasingly difficult. A number of
`situations require an identification of a person in our society: have I seen this
`applicant before? Is this person an employee of this company? 15 this individual a
`citizen of this country? Many situations will even warrant identification of a person at
`the far end of a communication channel.
`
`2. Opportunities
`
`Accurate identification of a person could deter crime and fraud, streamline business
`processes, and save critical resources. Here are a few mind boggling numbers: about
`$1 billion dollars in welfare benefits in the United States are annually claimed by
`“double dipping” welfare recipients with fraudulent multiple identities
`[10].
`MasterCard estimates the credit card fraud at $450 million per annum which includes
`charges made on lost and stolen credit cards: unobtrusive positive personal
`identification of the legitimate ownership of a credit card at the point of sale would
`greatly reduce the credit card fraud; about 1 billion dollars worth of cellular telephone
`calls are made by the cellular bandwidth thieves - many of which are made from
`stolen pins and/or cellular telephones. Again, an identification of ~the legitimate
`ownership of the cellular telephones would prevent cellular telephone thieves from
`stealing the bandwidth. A reliable method of authenticating legitimate owner of an
`ATM card would greatly reduce ATM related fi'aud worth approximately $3 billion
`annually [11]. A positive method of identifying the rightful check payee would also
`reduce billions of dollars misappropriated through fraudulent encashment of checks
`each year- A method of positive authentication of each system login would eliminate
`illegal break-ins into traditionally secure (even federal government) computers. The
`United States Immigration and Naturalization service stipulates that it could each day
`detect/deter about 3,000 illegal
`immigrants crossing the Mexican border without
`delaying the legitimate people entering the United States if it had a quick way of
`establishing positive personal identification.
`
`3.
`
`Identification Methods
`
`In a broad
`The problem of authentication and identification is very challenging.
`sense, establishing an identity (either in a verification context or an identification
`context) is a very difficult problem; Gertrude Stein's [12] quote “rose is a rose is a
`rose is a rose” summarizes the essence of the difficulty of a positive identification
`
`

`

`Introduction to Biometrics
`
`3
`
`problem: an identity of a person is so much woven into the fabric of everything that a
`person represents and believes that the answers to the identity of a person transcend
`the scope of an engineering system and the solutions could (perhaps) only be sought
`in a philosophical realm. For example, can a brain-dead person be identified as her
`fully sane connteipart for authenticating an electronic fund transfer? Engineering
`approach to the (abstract) problem of authentication of a person's identity is to reduce
`it to the problem of authentication of a concrete entity related to the person (Figure
`1.1). Typically, these entities include (i) a person's possession (“something that you
`possess”), e.g., permit physical access to a building to all persons whose identity
`could be authenticated by possession of a key; (ii) person‘s knowledge of a piece of
`information (“something that you know”), e.g., permit login access to a system to a
`person who knows the user-id and a password associated with it. Some systems, e.g.,
`ATMs, use a combination of “something that you have” (ATM card) and “something
`that you know” (PIN) to establish an identity. The problem with the traditional
`approaches of identification using possession as a means of identity is that
`the
`possessions could be lost, stolen, forgotten, or misplaced. Further, once in control of
`the identifying possession, by definition, any other “unauthorized” person could abuse
`the privileges of the authorized user. The problem with using knowledge as an
`identity authentication mechanism is
`that
`it
`is difficult
`to remember
`the
`passwordsfPINs; easily recallable passwordsfPINs (e.g., pet's name, spouse's birthday)
`
`
`
`Figure 1.1 Prevalent methods of identification based on possession and knowledge:
`Keys. employee badge, driver license, ATM card, and credit card.
`
`could be easily guessed by the adversaries. It has been estimated that about 25% of
`the people using ATM cards write their ATM Pl'Ns on the ATM card [13],
`thereby
`defeating possession/knowledge combination as a means of identification. As a result,
`these techniques cannot distinguish between an authorized person and an impostor
`who acquires the knowledge/possession, enabling the access privileges of the
`authorized person.
`
`

`

`4
`
`Jain et a1.
`
`Yet another approach to positive identification has been to reduce the problem of
`identification to the problem of identifying physical characteristics of the person. The
`characteristics could be either a person's physiological traits, e.g., fingerprints, hand
`geometry, etc. or her behavioral characteristics, e.g., voice and signature. This method
`of identification of a person based on hisfher physiological/behavioral characteristics
`is called biometricsz. The primary advantage of such an identification method over
`the methods of identification utilizing “something that you possess” or “something
`that you know” approach is that a biometrics cannot be misplaced or forgotten; it
`represents a tangible component of “something that you are”. While biometric
`techniques are not an identification panacea, they, especially, when combined with
`the other methods of identification, are beginning to provide very powerful tools for
`problems requiring positive identification.
`
`4. Biometrics
`
`What biological measurements qualify to be a biometric? Any human physiological
`or behavioral characteristic could be a biometrics provided it has the following
`desirable properties [15]: (i) universality, which means that every person should have
`the characteristic, (ii) uniqueness, which indicates that no two persons should be the
`same in terms of the characteristic,
`(iii) permanence, which means that
`the
`characteristic should be invariant with time, and (iv) collectabilinz, which indicates
`that the characteristic can be measured quantitatively. In practice, there are some other
`important requirements [15,16]:
`(1’) performance, which refers to the achievable
`identification accuracy,
`the resource requirements
`to achieve an acceptable
`identification accuracy, and the working or environmental factors that affect the
`identification accuracy, (ii) acceptability, which indicates to what extent people are
`willing to accept the biometric system, and (in) circumvention, which refers to how
`easy it is to fool the system by fraudulent techniques.
`
`5. Biometrics Technology: Overview
`
`No single biometrics is expected to effectively satisfy the needs of all identification
`(authentication) applications.
`A number of biometrics have been proposed,
`researched, and evaluated for
`identification (authentication) applications. Each
`biometrics has its strengths and limitations; and accordingly, each biometric appeals
`to a particular identification (authentigation) application. A summary of the existing
`and burgeoning biometric technologies is described in this section.
`
`I Voice
`
`Voice is a characteristic of an individual [17]. However, it is not expected to be
`sufficiently unique to permit identification of an individual from a large database
`of identities (Figure 1.2). Moreover, a voice signal available for authentication is
`
`2 Note the distinction between the terms biometrics and biometry: biometry encompasses a much broader
`field involving application of statistics to biology and medicine [14].
`
`

`

`Introduction to Biometrics
`
`5
`
`typically degraded in quality by the microphone, communication channel, and
`digitizer characteristics. Before extracting features, the amplitude of the input
`signal may be normalized and decomposed into several band-pass frequency
`channels. The features extracted from each band may be either time-domain or
`frequency domain features. One of the most commonly used features is cepstral
`feature - which is a logarithm of the Fourier Transform of the voice signal in
`each band. The matching strategy mayetypically employ approaches based on
`hidden Markov model, vector quantization, or dynamic time warping [17]. Text—
`dependent speaker verification authenticates the identity of a subject based on a
`fixed predetermined phrase.
`Text—independent speaker verification is more
`difficult and verifies a speaker identity independent of the phrase. Language—
`independent speaker verification verifies a speaker identity irrespective of the
`language of the uttered phrase and is even more challenging.
`
`2000
`
`" 1500
`
`1000
`
`
`
`0.15
`
`0.2
`
`_l_.
`0.25
`
`—L._ —J...
`0.3
`0.35
`
`,._a.
`0.05
`
`_-.__1
`0.1
`
`_J
`0.4
`
`0.45
`
`Figure 1.2 Voice signal representing an utterance of the word “seven". X and Y axes
`represent time and signal amplitude, respectively.
`
`Voice capture is unobtrusive and voice print is an acceptable biometric in
`almost all societies. Some applications entail authentication of identity over
`telephone. In such situations, voice may be the only feasible biometric. Voice is
`a behavioral biometrics and is affected by a person's health (e.g., cold), stress,
`emotions, etc. To extract features which remain invariant in such cases is very
`difficult. Besides, some people seem to be extraordinarily skilled in mimicking
`others. A reproduction of an earlier recorded voice can be used to circumvent a
`voice authentication system in the remote unattended applications. One of the
`methods of combating this problem is to prompt the subject (whose identity is to
`be authenticated) to utter a different phrase each time.
`
`

`

`Infrared Facial and Hand Vein Thermograms
`
`Jain et a1.
`
`
`
`Figure 1.3 Identification based on facial thermograms [1]. The image is obtained
`by sensing the infrared radiations from the face of a person. The graylevel at each
`pixel is characteristic of the magnitude of the radiation.
`
`Human body radiates heat and the pattern of heat radiation is a characteristic of
`each individual body [18]. An infrared sensor could acquire an image indicating
`the heat emanating from different parts of the body (Figure 1.3). These images
`are called thermograms. The method of acquisition of the thermal
`image
`unobtrusively is akin to the capture of a regular (visible spectrum) photograph of
`the person. Any part of the body could be used for identification. The absolute
`values of the heat radiation are dependent upon many extraneous factors and are
`not completely invariant to the identity of an individual; the raw measurements of
`heat radiation need to be normalized, e. g., with respect to heat radiating from a
`landmark feature of the body.
`The technology could be used for covert
`identification solutions and could distinguish between identical twins.
`It is also
`claimed to provide enabling technology for identifying people under the influence
`of drugs: the radiation patterns contain signature of each narcotic drug [19]. A
`thennogram—based
`system may have
`to address
`sensing challenges
`in
`uncontrolled environments, where heat emanating surfaces in the vicinity of the
`body, e.g., room heaters and vehicle exhaust pipes, may drastically affect the
`image acquisition phase.
`Infiargd facial thermograms seem to be acceptable
`since their acquisition is a non—contact and non-invasive sensing technique.
`Identification systems using facial thermograms are commercially available
`[1]. A related technology using near infrared imaging [2] is used to scan the back
`of a clenched fist to determine hand vein structure (Figure 1.4).
`Infrared sensors
`are prohibitively expensive which is a factor inhibiting wide spread use of
`thermograms.
`
`

`

`Introduction to Biometrics
`
`7
`
`
`
`Figure 1.4 Identification based on hand veins [2]. An infrared image of the back of a
`clenched human fist. The structure of the vasculature could be used for identification.
`
`-
`
`Fingerprints
`
`Fingerprints are graphical flow—like ridges present on human fingers. Their
`formations depend on the initial conditions of the embryonic development and
`they are believed to be unique to each person (and each finger). Fingerprints are
`one of the most mature biometric technologies used in forensic divisions
`worldwide for criminal investigations and therefore, have a stigma of criminality
`associated with them. Typically, a fingerprint image is captured in one of two
`ways: (i) scanning an inked impression of a finger or (ii) using a live-scan
`fingerprint scanner (Figure 1.5).
`Major representations of the finger are based on the entire image, finger
`ridges, or salient features derived from the ridges (minutiae).
`Four basic
`approaches to identification based on fingerprint are prevalent: (i) the invariant
`properties of the gray scale profiles of the fingerprint image or a part thereof; (ii)
`global ridge patterns, also known as fingerprint classes; (iii) the ridge patterns of
`the fingerprints; (iv) fingerprint minutiae — the features resulting mainly from
`ridge endings and bifurcations.
`
`

`

`Jain et al.
`
`
`
`Figure 1.5 A fingerprint image could be captured from the inked impression of a
`finger or directly imaging a finger using frustrated total
`internal reflection
`technology. The former is called an inked fingerprint (a) and the latter is called a
`live-scan fingerprint (b).
`
`Face
`
`Face is one of the most acceptable biometrics because it is one of the most
`common method of identification which humans use in their visual interactions
`
`In addition, the method of acquiring face images is non—intrusive.
`(Figure 1.6).
`Two primary approaches to the identification based on face recognition are the
`following: (i) Transform approach [20, 21]: the universe of face image domain is
`represented using a set of orthonormal basis vectors. Currently, the most popular
`basis vectors are eigenfaces: each eigenface is derived from the covariance
`analysis of the face image population; two faces are considered to be identical if
`they are sufficiently “close” in the eigenface feature space. A number of variants
`of such an approach exist.
`(ii) Attribute—based approach [22]: facial attributes
`like nose, eyes, etc. are extracted from the face image and the invariance of
`geometric properties among the face landmark features is used for recognizing
`features.
`
`It is
`Facial disguise is of concern in unattended authentication applications.
`very challenging to develop face recognition techniques which can tolerate the
`effects of aging, facial expressions, slight variations in the imaging environment
`and variations in the pose of face ”with respect to camera (2D and 3D rotations)
`[23].
`
`Iris
`
`Visual texture of the human iris is determined by the chaotic morphogenetic
`processes during embryonic development and is posited to be unique for each
`person and each eye [24]. An iris image is typically captured using a non-contact
`imaging process (Figure 1.7). The image is obtained using an ordinary CCD
`camera with a resolution of 512 dpi. Capturing an iris image involves cooperation
`
`

`

`Introduction to Biometrics
`
`9
`
`from the user, both to register the image of iris in the central imaging area and to
`ensure that the iris is at a predetermined distance from the focal plane of the
`camera. A position-invariant constant length byte vector feature is derived from
`an annular part of the iris image based on its texture. The identification error rate
`using iris technology is believed to be extremely small and the constant length
`position invariant code permits an extremely fast method of iris recognition.
`
`
`
`Figure 1.6 Identification based on face is one of the most acceptable methods of
`biometric-based identification.
`
`
`
`Figure 1.7 Identification Based on Iris. The visual texture of iris could be used for
`positive person identification.
`
`

`

`lO
`
`0
`
`Eat
`
`Jain et a1.
`
`It is known that the shape of the ear and the structure of the cartilegenous tissue
`of the pinna are distinctive3. The features of an ear are not expected to be unique
`to each individual. The ear recognition approaches are based on matching
`vectors of distances of salient points on the pinna from a landmark location
`(Figure 1.8) on the ear [3]. No commercial systems are available yet and
`authentication of individual identity based on ear recognition is still a research
`topic.
`
`
`
`Figure 1.8 An image of an ear and the features used for ear-based identification [3].
`Feature vector consists of the distances of various salient locations on the pinna from
`a landmark location.
`
`.
`
`Gait
`
`Gait is the peculiar way one walks and is a complex spatio—temporal behavioral
`biometrics. Gait
`is not supposed to be unique to each individual, but
`is
`sufficiently characteristic to allow identity authentication. Gait is a behavioral
`biometric and may not stay invariant especially over a large period of time, due to
`large fluctuations of body weightfimajor shift in the body weight (e.g., waddling
`gait during pregnancy [25], major injuries involving joints or brain (e.g.,
`cerebellar lesions in Parkinson disease [25]), or due to inebriety (e.g., drunken
`gait [25]).
`Humans are quite adept at recognizing a person at a distance from his gait.
`Although, the characteristic gait of a human walk has been well researched in
`biomechanics community to detect abnormalities in loWer extremity joints, the
`
`3 Department of Immigration and Naturalization in the United States specifically
`requests photographs of individuals with clearly visible right ear.
`
`

`

`Introduction to Biometrics
`
`l 1
`
`use of gait for identification purposes is very recent. Typically, gait features are
`derived from an analysis of a video-sequence footage (Figure 1.9) of a walking
`person [26] and consist of characterization of several different movements of
`each articulate joint. Currently, there do not exist any commercial systems for
`performing gait—based authentication. The method of input acquisition for gait is
`not different from that of acquiring facial pictures, and hence gait may be an
`acceptable biometric. Since gait determination involves processing of video, it is
`compute and input intensive.
`
`
`
`Figure 1.9 Authentication based on gait typically uses a sequence of images of a
`walking person. One of the frames in the image sequence is illustrated here.
`
`I Keystroke Dynamics
`It is hypothesized that each person types on a keyboard in a characteristic way.
`This behavioral biometrics is not expected to be unique to each individual but it
`offers sufficient discriminatory information to permit identity authentication [27].
`Keystroke dynamics is a behavioral biometric; for some individuals, one may
`expect to observe a large variations from typical typing patterns. The keystrokes
`of a person using a system could be monitored unobtrusively as that person is
`keying in other information. Keystroke dynamic features are based On time
`durations between the keystrokes. Some variants of identity authentication use ‘
`features based on inter-key delays as well as dwell times — how long a person
`holds down a key. Typical matching approaches use a neural network architecture
`to associate identity with the keystroke dynamics features. Some commercial
`systems are already appearing in the market.
`a DNA
`
`DNA (DeoxyriboNucleic Acid) is the one—dimensional ultimate unique code for
`one's individuality — except for the fact that identical twins have the identical
`DNA pattern. It is, however, currently used mostly in the context of forensic
`applications for identification [4]. Three iSSues limit the utility of this biometrics
`for other applications: (i) contamination and sensitivity: it is easy to steal a piece
`
`

`

`12
`
`Jain et al.
`
`of DNA from an unsuspecting subject to be subsequently abused for an ulterior
`purpose; (ii) automatic real—time identification issues: the present technology for
`genetic matching is not geared for online unobtrusive identifications. Most of the
`human DNA is identical for the entire human species and only some relatively
`small number of specific locations (polymorphic loci) on DNA exhibit individual
`variation. These variations are manifested either in the number of repetitions of a
`block of base sequence (length polymorphism) or in the minor non—functional
`perturbations of the base sequence (sequence polymorphism) [70]. The processes
`involved in DNA based personal
`identification determine whether two DNA
`samples originate from the same/different individual(s) based on the distinctive
`signature at one or more polymorphic loci. A major component of these processes
`now exist
`in the form of cumbersome chemical methods (wet processes)
`requiring an expert's skills. There does not seem to be any effort directed at a
`complete automation of all the processes.(iii) privacy issues: information about
`susceptibilities of a person to certain diseases could be gained from the DNA
`pattern and there is a concern that
`the unintended abuse of genetic code
`information may result in discrimination in e.g., hiring practices.
`
`
`
`Figure 1.10 DNA is double helix structure made of four bases: Adenine (A), Thymine
`(T), Cytosine (C), and Guanine (G) [4]. The sequence of bases is unique to each
`individual (with the exception of identicai twins) and could be used for positive person
`identification.
`L
`
`Signature and Acoustic Emissions
`
`to be a characteristic of that
`The way a person signs her name is known
`individual (Figure 1.11). Although signatures requ