Human identification based on gait / by Mark S. Nixon, Tieniu Tan, Rama Chellappa.

Format:

Book

Author/Creator:

Nixon, Mark S.

Contributor:

Tan, Tieniu.

Chellappa, Rama

Elsie de Renzo Orlando Fund.

Series:

Kluwer international series on biometrics ; 4.

Kluwer international series on biometrics ; 4

Language:

English

Subjects (All):

Biometric identification.

Gait in humans.

Physical Description:

x, 187 pages : illustrations ; 25 cm.

Place of Publication:

New York ; [Great Britain] : Springer, [2006]

Summary:

Biometrics now affects many people's lives, and is the focus of much academic research and commercial development. Gait is one of the most recent biometrics, with its own unique advantages. Gait recognizes people by the way they walk and run, analyzes movement, which in turn, implies analyzing sequences of images.

This professional book introduces developments from the laboratories of very distinguished researchers within this relatively new area of biometrics and clearly establishes human gait as a biometric. Human Identification Based on Gait provides a snapshot of all the biometric work in human identification by gait (all major centers for research are indicated in this book). To complete the picture, studies are included from medicine, psychology and other areas wherein we find not only justification for the use of gait as a biometric, but also pointers to techniques and to analysis. Human Identification Based on Gait is designed for a professional audience, composed of researchers and practitioners in industry. This book is also suitable as a secondary text for graduate-level students in computer science.

Contents:

1.1 Biometrics and Gait 1

1.2 Contexts 2

1.2.1 Immigration and Homeland Security 2

1.2.2 Surveillance 2

1.2.3 Human ID at a Distance (HiD) Program 3

2 Subjects Allied to Gait 5

2.2 Literature 5

2.3 Medicine and Biomechanics 6

2.3.1 Basic Gait Analysis 6

2.3.2 Variation in Gait Covariate Factors 10

2.4 Psychology 12

2.5 Computer Vision-Based Human Motion Analysis 13

2.6 Other Subjects Allied to Gait 15

3 Gait Databases 17

3.1 Early Databases 17

3.1.1 UCSD Gait Data 17

3.1.2 Early Soton Gait Data 18

3.2 Current Databases 20

3.2.1 Overall Design Considerations 20

3.2.2 NIST/USF Database 21

3.2.3 Soton Database 22

Laboratory Layout 24

Outdoor Data Design Issues 27

Acquisition Set-up Procedure 29

Filming Issues 29

Recording Procedure 30

Ancillary Data 31

3.2.4 CASIA Database 32

3.2.5 UMD Database 33

4 Early Recognition Approaches 35

4.1 Initial Objectives and Constraints 35

4.2 Silhouette Based 35

4.3 Model Based 39

5 Silhouette-Based Approaches 45

5.2 Extending Shape Description to Moving Shapes 48

5.2.1 Area Masks 49

5.2.2 Gait Symmetry 51

5.2.3 Velocity Moments 53

5.2.4 Results 54

Recognition by Area Masks 55

Recognition by Symmetry 58

Recognition by Velocity Moments 61

5.2.5 Potency of Measurements of Silhouette 63

5.3 Procrustes and Spatiotemporal Silhouette Analysis 65

5.3.1 Automatic Gait Recognition Based on Procrustes Shape Analysis 65

5.3.2 Silhouette Detection and Representation for Procrustes Analysis 66

Silhouette Extraction 66

Representation of Silhouette Shapes 68

5.3.3 Procrustes Gait Feature Extraction and Classification 69

Procrustes Shape Analysis 69

Gait Signature Extraction 69

Similarity Measure and Classifier 70

5.3.4 Spatiotemporal Silhouette Analysis Based Gait Recognition 70

Spatiotemporal Feature Extraction 72

Feature Extraction and Classification 73

5.3.5 Experimental Results and Analysis 77

Procrustes Shape Analysis 77

Spatiotemporal Silhouette Analysis 82

5.4 Modeling, Matching, Shape and Kinematics 89

5.4.1 HMM Based Gait Recognition 89

Gait Recognition Framework 90

Direct Approach 91

Indirect Approach 93

5.4.2 DTW Based Gait Recognition 94

Gait Recognition Framework 96

5.4.3 Shape and Kinematics 97

Shape Analysis 97

Dynamical Models 98

5.4.4 Results 100

HMM Based Gait Recognition 100

DTW Based Gait Recognition 102

Shape and Kinematics 104

6 Model-Based Approaches 107

6.2 Planar Human Modeling 109

6.2.1 Modeling Walking and Running 109

6.2.2 Model-Based Extraction and Description 111

6.3 Kinematics-based People Tracking and Recognition in 3D Space 114

6.3.1 Model-based People Tracking using Condensation 114

Human Body Model 115

Learning Motion Model and Motion Constraints 117

Experiments and Discussions 125

6.4 Other Approaches 131

6.4.1 Structure by Body Parameters 132

6.4.2 Structural Model-based Recognition 132

7 Further Gait Developments 135

7.1 View Invariant Gait Recognition 135

7.1.1 Overview of the Algorithm 136

7.1.2 Optical flow based SfM approach 137

7.1.3 Homography based approach 138

7.1.4 Experimental Results 138

7.2 Gait Biometric Fusion 141

7.3 Fusion of Static and Dynamic Body Biometrics for Gait Recognition 144

7.3.1 Overview of Approach 144

7.3.2 Classifiers and Fusion Rules 145

7.3.3 Experimental Results and Analysis 146

8 Future Challenges 151

Literature 157

Medicine and Biomechanics 157

Covariate factors 158

Psychology 159

Computer Vision-Based Analysis of Human Motion 160

Databases 161

Early work 162

Current approaches 163

Further Analysis 166

Other Related Work 169

Appendix 9.1 Southampton Data Acquisition Forms 171

Appendix 9.1.1 Laboratory Set-up Forms 171

Appendix 9.1.2 Camera Set-up Forms 175

Appendix 9.1.3 Session Coordinator's Instructions 180

Appendix 9.1.4 Subject Information Form 182.

Notes:

Includes bibliographical references (pages [157]-170) and index.

Local Notes:

Acquired for the Penn Libraries with assistance from the Elsie de Renzo Orlando Fund.

ISBN:

0387244247

0387294880

OCLC:

62307441

Publisher Number:

9780387244242

9780387294889 (e-book)