Visual Analysis of Behaviour

by Gong, Shaogang; Xiang, Tao

• ISBN13: 9780857296696
• ISBN10: 0857296698
• Format: Hardcover
• Copyright: 2011-06-09
• Publisher: Springer-Verlag New York Inc

Summary

This book presents a comprehensive treatment of visual analysis of behaviour from computational-modelling and algorithm-design perspectives. Topics: covers learning-group activity models, unsupervised behaviour profiling, hierarchical behaviour discovery, learning behavioural context, modelling rare behaviours, and #xE2;#xAC;Sman-in-the-loop#xE2;#xAC; active learning; examines multi-camera behaviour correlation, person re-identification, and #xE2;#xAC;Sconnecting-the-dots#xE2;#xAC; for abnormal behaviour detection; discusses Bayesian information criterion, Bayesian networks, #xE2;#xAC;Sbag-of-words#xE2;#xAC; representation, canonical correlation analysis, dynamic Bayesian networks, Gaussian mixtures, and Gibbs sampling; investigates hidden conditional random fields, hidden Markov models, human silhouette shapes, latent Dirichlet allocation, local binary patterns, locality preserving projection, and Markov processes; explores probabilistic graphical models, probabilistic topic models, space-time interest points, spectral clustering, and support vector machines.

Author Biography

Dr. Shaogang Gong is a Professor of Visual Computation in the School of Electronic Engineering and Computer Science at Queen Mary University of London, UK. Dr. Tao Xiang is a Lecturer at the same institution.

Table of Contents

Introduction
About Behaviour	p. 3
Understanding Behaviour	p. 4
Representation and Modelling	p. 5
Detection and Classification	p. 6
Prediction and Association	p. 6
Opportunities	p. 7
Visual Surveillance	p. 8
Video Indexing and Search	p. 8
Robotics and Healthcare	p. 9
Interaction, Animation and Computer Games	p. 9
Challenges	p. 10
Complexity	p. 10
Uncertainty	p. 10
The Approach	p. 11
References	p. 12
Behaviour in Context	p. 15
Facial Expression	p. 15
Body Gesture	p. 16
Human Action	p. 19
Human Intent	p. 21
Group Activity	p. 22
Crowd Behaviour	p. 24
Distributed Behaviour	p. 25
Holistic Awareness: Connecting the Dots	p. 28
References	p. 29
Towards Modelling Behaviour	p. 39
Behaviour Representation	p. 39
Object-Based Representation	p. 39
Part-Based Representation	p. 43
Pixel-Based Representation	p. 44
Event-Based Representation	p. 46
Probabilistic Graphical Models	p. 48
Static Bayesian Networks	p. 50
Dynamic Bayesian Networks	p. 50
Probabilistic Topic Models	p. 52
Learning Strategies	p. 53
Supervised Learning	p. 53
Unsupervised Learning	p. 54
Semi-supervised Learning	p. 55
Weakly Supervised Learning	p. 56
Active Learning	p. 56
References	p. 58
Sing-Object Behaviour
Understanding Facial Expression	p. 69
Classification of Images	p. 69
Local Binary Patterns	p. 70
Designing Classifiers	p. 73
Feature Selection by Boosting	p. 76
Manifold and Temporal Modelling	p. 78
Locality Preserving Projections	p. 78
Bayesian Temporal Models	p. 85
Discussion	p. 88
References	p. 90
Modelling Gesture	p. 95
Tracking Gesture	p. 95
Motion Moment Trajectory	p. 95
2D Colour-Based Tracking	p. 97
Bayesian Association	p. 99
3D Model-Based Tracking	p. 109
Segmentation and Atomic Action	p. 115
Temporal Segmentation	p. 115
Atomic Actions	p. 117
Markov Processes	p. 119
Affective State Analysis	p. 123
Space-Time Interest Points	p. 124
Expression and Gesture Correlation	p. 126
Discussion	p. 128
References	p. 128
Action Recognition	p. 133
Human Silhouette	p. 134
Hidden Conditional Random Fields	p. 135
HCRF Potential Function	p. 138
Observable HCRF	p. 138
Space-Time Clouds	p. 141
Clouds of Space-Time Interest Points	p. 141
Joint Local and Global Feature Representation	p. 149
Localisation and Detection	p. 150
Tracking Salient Points	p. 151
Automated Annotation	p. 154
Discussion	p. 157
References	p. 158
Group Behaviour
Supervised Learning of Group Activity	p. 163
Contextual Events	p. 164
Seeding Event: Measuring Pixel-Change-History	p. 164
Classification of Contextual Events	p. 167
Activity Segmentation	p. 170
Semantic Content Extraction	p. 170
Semantic Video Segmentation	p. 175
Dynamic Bayesian Networks	p. 181
Correlations of Temporal Processes	p. 181
Behavioural Interpretation of Activities	p. 185
Discussion	p. 190
References	p. 191
Unsupervised Behaviour Profiling	p. 193
Off-line Behaviour Profile Discovery	p. 194
Behaviour Patterns	p. 194
Behaviour Profiling by Data Mining	p. 195
Behaviour Affinity Matrix	p. 196
Eigendecomposition	p. 196
Model Order Selection	p. 197
Quantifying Eigenvector Relevance	p. 198
On-line Anomaly Detection	p. 200
A Composite Behaviour Model	p. 200
Run-Time Anomaly Measure	p. 203
On-line Likelihood Ratio Test	p. 204
On-line Incremental Behaviour Modelling	p. 205
Model Bootstrapping	p. 206
Incremental Parameter Update	p. 208
Model Structure Adaptation	p. 210
Discussion	p. 211
References	p. 212
Hierarchical Behaviour Discovery	p. 215
Local Motion Events	p. 216
Markov Clustering Topic Model	p. 217
Off-line Model Learning by Gibbs Sampling	p. 219
On-line Video Saliency Inference	p. 221
On-line Video Screening	p. 223
Model Complexity Control	p. 226
Semi-supervised Learning of Behavioural Saliency	p. 228
Discussion	p. 230
References	p. 231
Learning Behavioural Context	p. 233
Spatial Context	p. 235
Behaviour-Footprint	p. 235
Semantic Scene Decompostion	p. 236
Correlational and Temporal Context	p. 238
Learning Regional Context	p. 239
Learning Global Context	p. 243
Context-Aware Anomly Detection	p. 245
Discussion	p. 247
References	p. 248
Modelling Rare and Subtle Behaviours	p. 251
Weakly Supervised Joint Topic Model	p. 252
Model Structure	p. 252
Model Parameters	p. 255
On-line Behaviour Classification	p. 262
Localisation of Rare Behaviour	p. 263
Discussion	p. 264
References	p. 265
Man in the Loop	p. 267
Active Behaviour Learning Strategy	p. 269
Local Block-Based Behaviour	p. 271
Bayesian Classification	p. 273
Query Criteria	p. 274
Likelihood Criterion	p. 275
Uncertainty Criterion	p. 275
Adaptive Query Selection	p. 277
Discussion	p. 279
References	p. 282
Distributed Behaviour
Multi-camera Behaviour Correlation	p. 285
Multi-view Activity Representation	p. 288
Local Bivariate Time-Series Events	p. 288
Activity-Based Scene Decomposition	p. 289
Learning Pair-Wise Correlation	p. 292
Cross Canonical Correlation Analysis	p. 293
Time-Delayed Mutual Information Analysis	p. 295
Multi-camera Topology Inference	p. 296
Discussion	p. 298
References	p. 298
Person Re-identification	p. 301
Re-identification by Ranking	p. 303
Support Vector Ranking	p. 303
Scalability and Complexity	p. 305
Ensemble Rank SVM	p. 306
Context-Aware Search	p. 308
Discussion	p. 310
References	p. 312
Connecting the Dots
Global Behaviour Segmentation	p. 315
Bayesian Behaviour Graphs	p. 315
A Time-Delayed Probalistic Graphical Model	p. 319
Bayesian Graph Structure Learning	p. 321
Bayesian Graph Parameter Learning	p. 326
Cumulative Anomaly Score	p. 326
Incremental Model Structure Learning	p. 329
Global Awareness	p. 334
Time-Ordered Latent Dirichlet Allocation	p. 334
On-line Prediction and Anomaly Detection	p. 337
Discussion	p. 339
References	p. 340
Epilogue	p. 343
Index	p. 345
Table of Contents provided by Ingram. All Rights Reserved.