Signal and Image Processing for Biometrics E-Book

• Herausgeber: John Wiley & Sons
• Kategorie: Wissenschaft und neue Technologien
• Sprache: Englisch

Seitenzahl: 473

Veröffentlichungsjahr: 2012

Beschreibung

The aim of this book is to deal with biometrics in terms of signal and image processing methods and algorithms. This will help engineers and students working in digital signal and image processing deal with the implementation of such specific algorithms.

It discusses numerous signal and image processing techniques that are very often used in biometric applications. In particular, algorithms related to hand feature extraction, speech recognition, 2D/3D face biometrics, video surveillance and other interesting approaches are presented. Moreover, in some chapters, Matlab codes are provided so that readers can easily reproduce some basic simulation results.

This book is suitable for final-year undergraduate students, postgraduate students, engineers and researchers in the field of computer engineering and applied digital signal and image processing.

Leseprobe

Table of Contents

Preface

Chapter 1. Introduction to Biometrics

1.1. Background: from anthropometry to biometrics

1.2. Biometrics today

1.3. Different modes of use of a biometric system and associated uses

1.4. Biometrics as a pattern recognition problem

1.5. Evaluation of different modalities

1.6. Quality

1.7. Multimodality

1.8. Biometrics and preservation of privacy

1.9. Conclusion

1.10. Bibliography

Chapter 2. Introduction to 2D Face Recognition

2.1. Introduction

2.2. Global face recognition techniques

2.3. Local face recognition techniques

2.4. Hybrid face recognition techniques

2.5. Some guidances

2.6. Some databases

2.7. Conclusion

2.8. Bibliography

Chapter 3. Facial Soft Biometrics for Person Recognition

3.1. Introduction to soft biometrics

3.2. Soft biometric systems for human identification

3.3. Overall error probability of a soft biometrics system

3.4. Conclusions and future directions

3.5. Bibliography

Chapter 4. Modeling, Reconstruction and Trackingfor Face Recognition

4.1. Background

4.2. Types of available information

4.3. Geometric approaches for the reconstruction

4.4. Model-based approaches for reconstruction

4.5. Hybrid approaches

4.6. Integration of the time aspect

4.7. Conclusion

4.8. Bibliography

Chapter 5. 3D Face Recognition

5.1. Introduction

5.2. 3D face databases

5.3. 3D acquisition

5.4. Preprocessing and normalization

5.5. 3D face recognition

5.6. Asymmetric face recognition

5.7. Conclusion

5.8. Bibliography

Chapter 6. Introduction to Iris Biometrics

6.1. Introduction

6.2. Iris biometric systems

6.3. Iris recognition methods: state-of-the-art

6.4. Preprocessing of iris images

6.5. Features extraction and encoding

6.6. Similarity measure between two IrisCodes

6.7. Iris biometrics: emerging methods

6.8. Conclusion

6.9. Bibliography

Chapter 7. Voice Biometrics: Speaker Verification and Identification

7.1. Introduction

7.2. Acoustic analysis for robust speaker recognition

7.3. Distributed speaker recognition through UBM—GMM models

7.4. Performance evaluation of DSIDV

7.5. Conclusion

7.6. Bibliography

Chapter 8. Introduction to Hand Biometrics

8.1. Introduction

8.2. Characterization by minutiae extraction

8.3. A few databases

8.4. Conclusion

8.5. Bibliography

Chapter 9. Multibiometrics

9.1. Introduction

9.2. Different principles of multibiometrics

9.3. Fusion levels

9.4. Applications and illustrations

9.5. Conclusion

9.6. Bibliography

Chapter 10. Hidden Biometrics

10.1. Introduction

10.2. Biometrics using ECG

10.3. Biometrics using EMG: preliminary experiments

10.4. Biometrics using medical imaging

10.5. Conclusion

10.6. Bibliography

Chapter 11. Performance Evaluation of Biometric Systems

11.1. Introduction

11.2. Reminders on biometric systems

11.3. Results analysis tools

11.4. Illustration of the GREYC-Keystroke system

11.5. Conclusion

11.6. Bibliography

Chapter 12. Classification Techniques for Biometrics

12.1. Introduction

12.2. Generalization aptitude and performance measures

12.3. Parametric approaches

12.4. Non-parametric approaches

12.5. Conclusion

12.6. Bibliography

Chapter 13. Data Cryptography

13.1. Introduction

13.2. Cryptography

13.3. Conclusion

13.4. Bibliography

Chapter 14. Visual Data Protection

14.1. Introduction

14.2. Visual data hiding

14.3. A proposed homomorphism-based visual secret sharing scheme

14.4. Conclusion

14.5. Bibliography

Chapter 15. Biometrics in Forensics

15.1. Introduction

First published 2012 in Great Britain and the United States by ISTE Ltd and John Wiley & Sons, Inc.

Apart from any fair dealing for the purposes of research or private study, or criticism or review, as permitted under the Copyright, Designs and Patents Act 1988, this publication may only be reproduced, stored or transmitted, in any form or by any means, with the prior permission in writing of the publishers, or in the case of reprographic reproduction in accordance with the terms and licenses issued by the CLA. Enquiries concerning reproduction outside these terms should be sent to the publishers at the undermentioned address:

ISTE Ltd 27-37 St George’s Road London SW19 4EU UK

www.iste.co.uk

John Wiley & Sons, Inc. 111 River Street Hoboken, NJ 07030 USA

www.wiley.com

© ISTE Ltd 2012

The rights of Amine Naït-Ali & Régis Fournier to be identified as the author of this work have been asserted by them in accordance with the Copyright, Designs and Patents Act 1988.

Library of Congress Cataloging-in-Publication Data

Signal and image processing for biometrics / edited by Amine Na t-Ali, Régis Fournier. p. cm. Includes bibliographical references and index.     ISBN 978-1-84821-385-2   1. Biometric identification. 2. Image processing. 3. Signal processing. I. Naït-Ali, Amine. II. Fournier, Régis.     TK7882.B56S54 2012     570.1'5195--dc23

2012017918

British Library Cataloguing-in-Publication Data

Preface

Literally, the word “biometrics” is composed of the prefix “bio”, meaning “life” in Greek, and the suffix “metrics”, meaning “measure”. Indeed, its main aim is to perform measurements on human beings. The term also refers to a discipline describing statistical and mathematical methods employed to process data related to life sciences. Even though several scientific communities “share” the word “biometrics” to deal with some specific fields such as medicine and ecology, this book is dedicated basically for security purposes. Actually, we believe that it is within this context that biometrics is best known, including the non-scientific community. This biometrics considers, in particular, the problem of identification and authentication of individuals using their characteristics. In fact, this issue goes back to the 19th Century, especially through the practice of anthropometry, but it has probably existed, under other forms, long before that time. The history of biometrics is exciting but its evolution is even more exciting. Over generations, human attitudes have changed and continued to mutate to the point where the degree of acceptability and subtlety is constantly evolving depending on the needs, constraints, and events that the world is aware of. Biometrics has changed a lot, and this term is becoming more and more common in our everyday language: the biometric passport, the biometric ID card, the biometric lock, etc. In the coming years, we expect that huge applications dealing with biometric-based systems will be developed. For example, according to the analysis of the biometrics market, published by the “International Biometric Group (IBG)”, we can find an increasing trend for both businesses and public sectors. As an example, we may mention one of the major programs of biometric identification that no country has ever known before. It consists of enrolling all Indian citizens to build a single national database.

Potentially, biometrics can be considered as an effective measure to allow an ease-of-use of technical systems or to provide some solutions to socioeconomic, management, and security issues. However, it is important to emphasize the fact that biometrics should be taught and controlled so that human identity, privacy, and freedom can be respected and that ethics is a priority or even a fundamental condition for the balance of the contemporary society.

Technically speaking, biometrics, as considered so far, would certainly not have existed without the progress reached in other disciplines, such as electronics, computer science, and signal and image processing. Within this context, many excellent books on biometrics have been published in recent years, highlighting both the software and the hardware aspects and considering more specifically acquisition systems and data processing techniques. But, this book is somehow different in the sense that the purpose is basically to provide a survey on biometrics as represented by French and some French-speaking research teams. The aim is to help postgraduate students, researchers, and engineers who need an introduction to biometrics and those who want to major in this field. In addition, we have tried to strike a balance between the chapters dedicated to research and those proposed for educational purpose by including Matlab code.

As the book title suggests, signal and image processing methods are presented by considering applications related to the identification and authentication of individuals. Obviously, two-dimensional/three-dimensional (2D/3D) face recognition, iris, and hand biometrics are considered, but the contents of this book are also extended to multibiometrics as well as to the performance evaluation of biometric systems. In addition, some signal processing tools such as classification, cryptography, and data protection are also presented.

The book consists of 15 chapters and is structured as follows:

Chapter 1: entitled “Introduction to Biometrics”. The history of biometrics is briefly reviewed; then the most common biometric modalities and their evaluation are presented. The multimodality and the privacy aspects are also discussed.

Chapter 2: in this chapter, “Introduction to 2D Face Recognition”, is proposed for educational purposes. It is especially intended for beginners. Its aim is to introduce some classical techniques and algorithms of facial biometrics by considering some local, global, and hybrid approaches.

Chapter 3: in this chapter, entitled “Facial Soft Biometrics for Person Recognition”, the aim is to deal with a specific type of biometrics that deals with some traits, such as the color of the eyes and hair, to identify persons or a group of persons.

Chapter 4: entitled “Modeling, Reconstruction and Tracking for Face Recognition”, the chapter addresses issues related to the acquisition of faces “on the fly”, in particular, by the use of multiview acquisition systems. Within an authentication context, the issues related to the 3D shape and to the texture of the face are addressed.

Chapter 5: in this chapter, entitled “3D Face Recognition”, 3D acquisition for biometrics, the preprocessing, as well as the symmetric and asymmetric face recognition are discussed.

Chapter 6: biometrics cannot be presented without addressing the iris modality. This is indeed the purpose of this chapter entitled “Introduction to Iris Biometrics”. The overall architecture of an iris biometric system is presented, which is essentially helpful for beginners. Afterwards, a step-by-step reference processing technique is detailed.

Chapter 7: in this chapter, entitled “Voice Biometrics: Speaker Verification and Identification”, some signal processing tools, in particular those used for analysis, modeling, and filtering, are elaborated within the context of speech recognition.

Chapter 8: this is another chapter for beginners, entitled “Introduction to Hand Biometrics”, in which the reader can perform basic processing (e.g. minutia extraction), using Matlab code. Links to several helpful databases are also provided.

Chapter 9: entitled “Multibiometrics”, this chapter presents various structures of multibiometric systems and the different biometric data fusion methods. Illustrations derived from industrial systems are also presented.

Chapter 10: in this chapter, biometrics is seen from a different viewpoint, in comparison with common techniques. Specifically, it consists of extracting signatures from biosignals and medical images for the purpose of identification or authentication. This biometrics, which is particularly robust to “spoofing”, is called “hidden biometrics”. In particular, we focus on biometrics using electrocardiogram (ECG), electromyogram (EMG), and some medical imaging techniques (e.g. brain MRI images, hand X-ray images, and anatomic images).

Chapter 11: after a brief review of some common definitions related to biometric systems, this chapter, entitled “Performance Evaluation of Biometric Systems”, is dedicated to the presentation of some tools used to assess the performance of biometric systems. Furthermore, interesting illustrations on keystroke dynamics systems are also presented.

Chapter 12: in biometric applications, it is often necessary to use classification techniques to associate a given feature with a predefined class. For this purpose, we present in this chapter, entitled “Classification Techniques for Biometrics”, numerous parametric (e.g. naive Bayesian and linear discriminant analysis (LDA)) and non-parametric (e.g. k-nearest neighbor (KNN), neural networks, and support vector machine (SVM)) methods. Matlab codes are also included.

Chapter 13: the main purpose of this chapter entitled “Data Cryptography” is to understand the basics of cryptography, including modern cryptography. It is obvious that the reader can use such a tool to encrypt biometric data.

Chapter 14: this chapter, entitled “Visual Data Protection”, is complementary to the previous chapter. It is dedicated to the protection of visual data through some specific methods, including digital watermarking and fingerprinting.

Chapter 15: in this chapter, entitled “Biometrics in Forensics”, the issues of facial comparison and voice comparison are discussed within the forensic context. The inference of the identity in forensics is also considered.

Finally, it is important to point out that this book would not have been possible without the active contribution of researchers from the French and French-speaking biometric community as well as some non-French-speaking researchers. This book is also the result of the participation of some members representing major industries and institutions active in the fields of biometrics and security. It is to all these participants that we wish to express our gratitude.

Amine NAIT-ALI and Régis FOURNIERJune 2012

Chapter 1Introduction to Biometrics1

Nowadays, biometrics is an emerging technique that allows us to verify the identity of an individual by using one or more of his or her personal characteristics. Its advantage is to increase the level of security by using as an identifier data that cannot be lost, stolen, or tampered with unlike passwords or personal identification number (PIN) codes, since they are directly related to the body or the behavior of the individual. A resurgence of interest in these techniques has been observed since the 2000s, a period when security policies were implemented in the G8 countries following the attacks of 9/11, among others. Recently several big deployments of biometrics systems have taken place. Let us quote the biometric passport, national identity cards and the new census of the Indian population. The purpose of this chapter is to give a brief introduction to biometric systems and to the various challenges that remain to be tackled by researchers of the field, in particular to cope with these large-scale deployments.

1.1. Background: from anthropometry to biometrics

Biometrics first emerged in the late 19th Century for police usage only. The taking of fingerprints, which is the oldest of biometric technologies, ultimately prevailed in the 19th Century for the identification of individuals, including criminals, after the work of various anthropologists, notably the English anthropologist Francis Galton in 1892.

In France, around 1880, Alphonse Bertillon developed forensic science through the implementation of anthropometric data sheets for each arrested person. The data sheets were used to identify detainees using the metric survey of their anatomical characteristics. This method gave him worldwide success but hid from him the global progress of dactyloscopy. At last, he agreed to add fingerprints to his data sheets. Then, in 1902, he identified the perpetrator of a crime through his fingerprints (Scheffer case) after failing with an anthropometry test [SCI 10].

1.2. Biometrics today

Different biometric modalities have been published. We distinguish between physiological modalities (iris, fingerprints, hand veins, etc.) that are more stable over time, a priori, and can be acquired with much difficulty, and behavioral modalities (handwriting, gait, keystroke dynamics, etc.) that are not only more variable but also more natural and can be acquired through simple and user-friendly means. Biological modalities can also be used (cardiac signal, see Chapter 10, DNA). They are more difficult to process for an immediate identification.

Nowadays, it is possible to process biometric data using a computer because we can digitize, store, and retrieve them from databases. This may thus lead to large-scale deployments, which are only made possible because of this “digitization” of personal and corporal information.