Biometric Inverse Problems

by Yanushkevich; Svetlana N.

• ISBN13: 9780849328992
• ISBN10: 0849328993
• Format: Hardcover
• Copyright: 2005-05-05
• Publisher: CRC Press

Summary

Traditional methods of biometric analysis are unable to overcome the limitations of existing approaches, mainly due to the lack of standards for input data, privacy concerns involving use and storage of actual biometric data, and unacceptable accuracy. Exploring solutions to inverse problems in biometrics transcends such limits and allows rich analysis of biometric information and systems for improved performance and testing. Although some particular inverse problems appear in the literature, until now there has been no comprehensive reference for these problems.Biometric Inverse Problems provides the first comprehensive treatment of biometric data synthesis and modeling. This groundbreaking reference comprises eight self-contained chapters that cover the principles of biometric inverse problems; basics of data structure design; new automatic synthetic signature, fingerprint, and iris design; synthetic faces and DNA; and new tools for biometrics based on Voronoi diagrams. Based on the authors' vast experience in the field, the book authoritatively examines new approaches and methodologies in both direct and inverse biometrics, providing invaluable analytical and benchmarking tools. The authors include case studies, examples, and implementation codes for practical illustration of the methods.Loaded with approximately 200 figures, 60 problems, 50 MATLAB® code fragments, and 200 examples, Biometric Inverse Problems sets the standard for innovation and authority in biometric data synthesis, modeling, and analysis.

Table of Contents

Preface	xiii
Acknowledgments	xxv
1 Introduction to the Inverse Problems of Biometrics	1	(56)
1.1 Attacks on biometric systems	3	(7)
1.1.1 Forgery of biometric data	3	(1)
1.1.2 Scenario of attacks	4	(2)
1.1.3 Examples of forgeries	6	(2)
1.1.4 Trends	8	(2)
1.2 Classical direct and inverse problems	10	(3)
1.3 Direct and inverse problems of biometrics	13	(1)
1.3.1 Automated analysis of biometric information	13	(1)
1.3.2 Automated synthesis of biometric information	14	(1)
1.4 Basic notion of biometric data	14	(6)
1.4.1 Representation of biometric data	15	(1)
1.4.2 Description of biometric data	16	(1)
1.4.3 Design of synthetic biometric data	17	(2)
1.4.4 Artificial intelligence paradigm in synthetic biometric data design	19	(1)
1.5 Examples of synthetic biometric data	20	(5)
1.6 Conversating of biometric information	25	(3)
1.6.1 Types of conversions	25	(1)
1.6.2 Brain-machine interface	26	(2)
1.6.3 Implementation	28	(1)
1.7 Design of biometric devices and systems	28	(5)
1.7.1 Functions	29	(1)
1.7.2 Design styles for various applications	30	(1)
1.7.3 Architecture	30	(2)
1.7.4 Databases	32	(1)
1.8 Applications of inverse biometrics	33	(3)
1.8.1 Training	33	(1)
1.8.2 Intellectual property privacy protection	34	(2)
1.9 Ethical and social aspects of inverse biometrics	36	(1)
1.9.1 Roles for practical ethics linked to social sciences	36	(1)
1.9.2 Prevention of undesirable side effects	36	(1)
1.9.3 Targeting of areas of social concern in biometrics	37	(1)
1.10 Summary	37	(2)
1.11 Problems	39	(1)
1.12 Further reading	40	(8)
1.13 References	48	(9)
2 Basics of Synthetic Biometric Data Structure Design	57	(48)
2.1 Basic concepts of synthetic biometric data structure design	58	(5)
2.1.1 Topology	58	(1)
2.1.2 Model	58	(3)
2.1.3 Analysis-by-synthesis modeling	61	(1)
2.1.4 Acceptable and unacceptable data	62	(1)
2.2 Synthesis strategies	63	(4)
2.2.1 Assembling	63	(2)
2.2.2 Design based on modification of biometric data	65	(1)
2.2.3 Functional assembling and physical modeling	66	(1)
2.2.4 Testing	67	(1)
2.3 Information carried by biometric data	67	(1)
2.3.1 Categories of information	67	(1)
2.3.2 Measuring of information	68	(1)
2.4 Generation of random biometric attributes	68	(6)
2.4.1 The simplest noise models	70	(1)
2.4.2 Generation of random numbers with prescribed distributions	71	(3)
2.5 Degradation model in synthesis of biometric data	74	(2)
2.5.1 Direct and inverse models of image degradation	74	(1)
2.5.2 Restoration of an image containing additive noise	75	(1)
2.6 Image warping	76	(4)
2.6.1 Image warping as a topological transformation	76	(1)
2.6.2 Spatial transforms	77	(1)
2.6.3 Controlled distortion by affine transform	78	(2)
2.7 Deformation by interpolation and approximation	80	(5)
2.7.1 Parametric and implicit models	80	(4)
2.7.2 Generating curves by approximation	84	(1)
2.8 Extracting and generating features	85	(4)
2.8.1 Features	85	(2)
2.8.2 Spatial feature extraction and generation	87	(2)
2.9 Summary	89	(1)
2.10 Problems	90	(6)
2.11 Further reading	96	(5)
2.12 References	101	(4)
3 Synthetic Signatures	105	(42)
3.1 Introduction	106	(3)
3.1.1 Off-line and on-line automated signature processing	106	(2)
3.1.2 Information carried by signatures	108	(1)
3.1.3 Signature acquisition	108	(1)
3.2 Basics of signature synthesis	109	(4)
3.2.1 Signature geometry	109	(2)
3.2.2 Requirements for signature synthesis	111	(1)
3.2.3 Scenarios for signature generation	112	(1)
3.3 Signature synthesis techniques	113	(3)
3.3.1 Deformation technique	114	(1)
3.3.2 Assembling technique	114	(1)
3.3.3 Cancellable technique	115	(1)
3.4 Statistically meaningful synthesis	116	(13)
3.4.1 Preprocessing: learning	117	(1)
3.4.2 Random number generation	118	(1)
3.4.3 Geometrically meaningful synthesis	118	(7)
3.4.4 Kinematically meaningful synthesis	125	(4)
3.4.5 Postprocessing: validation	129	(1)
3.5 Implementation	129	(6)
3.6 Summary	135	(1)
3.7 Problems	136	(3)
3.8 Further reading	139	(3)
3.9 References	142	(5)
4 Synthetic Fingerprints	147	(48)
4.1 Introduction	148	(4)
4.2 Modeling a fingerprint	152	(3)
4.2.1 Topology	152	(1)
4.2.2 Information carried by the fingerprint	153	(2)
4.3 Extraction of features	155	(1)
4.4 Library of topological primitives	156	(1)
4.4.1 Terminology	157	(1)
4.4.2 Topological primitive and macroprimitive	157	(1)
4.5 Local generators: fingerprint primitives	157	(3)
4.5.1 The simplest generator of primitives based on the curve model	159	(1)
4.5.2 The simplest generator of primitives based on the spline model	159	(1)
4.6 Global generators: orientation map	160	(4)
4.6.1 Parameters of orientation map	161	(3)
4.6.2 Direct orientation transform	164	(1)
4.7 Polar transform of orientation map	164	(6)
4.7.1 Extraction of reference and center point	165	(2)
4.7.2 Polar transform	167	(1)
4.7.3 Creation of a new orientation map	168	(2)
4.8 Generating synthetic fingerprint images from an orientation map	170	(4)
4.9 Other global topological models	174	(4)
4.9.1 Master fingerprint	175	(1)
4.9.2 Simulation of noise and other conditions	176	(2)
4.10 Summary	178	(3)
4.11 Problems	181	(4)
4.12 Further reading	185	(4)
4.13 References	189	(6)
5 Synthetic Faces	195	(66)
5.1 Introduction to facial expressions design	196	(3)
5.1.1 Terminology	196	(1)
5.1.2 Applications	197	(1)
5.1.3 Modeling	197	(1)
5.1.4 Direct and inverse problems	198	(1)
5.1.5 Data structures of facial synthesis	198	(1)
5.2 Modeling of facial expressions	199	(10)
5.2.1 Basic properties of facial expressions	199	(3)
5.2.2 Components of a face model	202	(1)
5.2.3 Facial synthesis based on physical models	202	(2)
5.2.4 Facial behavior models	204	(2)
5.2.5 Library	206	(2)
5.2.6 Distribution of information sources across an object	208	(1)
5.3 Facial topology transformation and manipulation	209	(4)
5.3.1 Automated facial measurements	210	(1)
5.3.2 Morphing technique	211	(1)
5.3.3 Facial deformation	211	(1)
5.3.4 Local and global spatial transformations	211	(2)
5.4 Local facial models	213	(8)
5.4.1 Information carried by eyes, eyebrows, and lips	213	(1)
5.4.2 Eye synthesis	214	(1)
5.4.3 Eyebrow synthesis	215	(1)
5.4.4 Mouth and lips synthesis	216	(3)
5.4.5 Facial caricatures	219	(2)
5.5 Facial synthesizers	221	(4)
5.5.1 Speech-to-lips systems	222	(1)
5.5.2 Facial transmission	223	(2)
5.6 Automated support of deceit detection	225	(8)
5.6.1 Lie detection problem	225	(1)
5.6.2 Face as a carrier of false information	226	(2)
5.6.3 Models of controlled and non-controlled facial dynamics	228	(2)
5.6.4 New generation of lie detectors	230	(3)
5.7 Summary	233	(1)
5.8 Problems	234	(3)
5.9 Further reading	237	(15)
5.10 References	252	(9)
6 Synthetic Irises	261	(36)
6.1 State-of-the-art iris synthesis	262	(2)
6.1.1 Non-automated design technique	262	(1)
6.1.2 Automated design technique	263	(1)
6.2 Eye model	264	(4)
6.2.1 Structure of eye and models	265	(1)
6.2.2 Structure of iris and models	265	(3)
6.2.3 Measures	268	(1)
6.2.4 Information carried by the iris	268	(1)
6.3 Iris image preprocessing	268	(4)
6.3.1 Iris localization	269	(1)
6.3.2 Iris normalization	270	(1)
6.3.3 Iris image enhancement and noise removal	270	(1)
6.3.4 Features in analysis and synthesis	271	(1)
6.4 Iris synthesis by transformation	272	(1)
6.4.1 Recombination of iris images	272	(1)
6.4.2 Iris image warping	272	(1)
6.5 Iris synthesis by assembling	272	(11)
6.5.1 Assembling	275	(1)
6.5.2 Master iris	275	(1)
6.5.3 Library of iris colors and topological primitives	276	(1)
6.5.4 Automated generation of iris primitives	276	(1)
6.5.5 Primitive template	277	(4)
6.5.6 Using Fourier transform in collarette design	281	(2)
6.5.7 Other transforms for automated design of iris primitives	283	(1)
6.6 Summary	283	(2)
6.7 Problems	285	(6)
6.8 Further reading	291	(3)
6.9 References	294	(3)
7 Biometric Data Structure Representation by Voronoi Diagrams	297	(52)
7.1 Voronoi data structure	298	(2)
7.1.1 Topology	298	(1)
7.1.2 Computing Voronoi diagram	299	(1)
7.2 Basics of Voronoi diagram technique	300	(8)
7.2.1 Metric space	300	(2)
7.2.2 Voronoi diagram	302	(5)
7.2.3 Information carried by a Voronoi diagram	307	(1)
7.3 Direct and inverse Voronoi transform	308	(5)
7.3.1 Preliminaries	308	(2)
7.3.2 Direct Voronoi transform	310	(1)
7.3.3 Inverse Voronoi transform	310	(1)
7.3.4 Extension	311	(2)
7.4 Properties	313	(6)
7.5 Voronoi data structure in topological analysis and synthesis	319	(3)
7.6 Topological compatibility of the Voronoi diagram	322	(4)
7.6.1 Distance mapping of feature points	322	(1)
7.6.2 Distance map	323	(3)
7.7 Implementing the discrete Voronoi transform with a distance transform	326	(5)
7.7.1 Distance mapping of biometric data	329	(1)
7.7.2 Transformation of the distance map of point generators to a discrete approximation of Voronoi diagrams	329	(2)
7.8 Calculating area Voronoi diagrams using nearest-neighbor transform	331	(4)
7.8.1 Effects of approximation to Voronoi diagrams via distance transform	331	(2)
7.8.2 Nearest-neighbor map	333	(1)
7.8.3 Transformation of nearest-neighbor map of area primitives into a Voronoi diagram	334	(1)
7.9 Summary	335	(3)
7.10 Problems	338	(6)
7.11 Further reading	344	(2)
7.12 References	346	(3)
8 Synthetic DNA	349	(30)
8.1 Introduction	351	(2)
8.1.1 DNA characteristics	351	(1)
8.1.2 Ideal identification properties	351	(1)
8.1.3 Privacy issues	352	(1)
8.2 Basics of DNA biometrics	353	(3)
8.2.1 DNA code	353	(1)
8.2.2 Terminology	354	(2)
8.3 DNA/protein synthesis techniques	356	(5)
8.3.1 Mutation technique	356	(1)
8.3.2 Recombination technique	357	(2)
8.3.3 Statistically-driven synthesis	359	(2)
8.4 Examples of Markov models	361	(4)
8.4.1 0th order Markov model	361	(1)
8.4.2 1st order Markov model	362	(1)
8.4.3 2nd order Markov model	363	(2)
8.5 Postprocessing: pairwise alignments	365	(4)
8.6 Algorithm for sequence generation	369	(1)
8.7 Summary	370	(1)
8.8 Problems	371	(3)
8.9 Further reading	374	(3)
8.10 References	377	(2)
Index	379