Statistical learning and pattern analysis for image and video processing / Nanning Zheng, Jianru Xue.

Format:

Book

Author/Creator:

Zheng, Nanning.

Contributor:

Xue, Jianru.

Series:

Advances in pattern recognition

Language:

English

Subjects (All):

Pattern recognition systems.

Image processing--Statistical methods.

Image processing.

Physical Description:

xvi, 365 pages : illustrations (some color) ; 24 cm.

Place of Publication:

Dordrecht ; New York : Springer, [2009]

Summary:

The inexpensive collection, storage, and transmission of vast amounts of visual data has revolutionized science, technology, and business. Innovations from various disciplines have aided in the design of intelligent machines able to detect and exploit useful patterns in data. One such approach is statistical learning for pattern analysis.

Among the various technologies involved in intelligent visual information processing, statistical learning and pattern analysis is undoubtedly the most popular and important approach, and is the area which has undergone the most rapid development in recent years. Above all, it provides a unifying theoretical framework for applications of visual pattern analysis.

This unique textbook/reference provides a comprehensive overview of theories, methodologies, and recent developments in the field of statistical learning and statistical analysis for visual pattern modeling and computing. The book describes the solid theoretical foundation, provides a complete summary of the latest advances, and presents typical issues to be considered in making a real system for visual information processing.

Features:

Provides a broad survey of recent advances in statistical learning and pattern analysis with respect to the two principal problems of representation and computation in visual computing

Presents the fundamentals of statistical pattern recognition and statistical learning via the general framework of a statistical pattern recognition system

Discusses pattern representation and classification, as well as concepts involved in supervised learning, semi-statistical learning, and unsupervised learning

Introduces the supervised learning of visual patterns in images, with a focus on supervised statistical pattern analysis, feature extraction and selection, and classifier design.

Covers visual pattern analysis in video, including methodologies for building intelligent video analysis systems, critical aspects of motion analysis, and multi-target tracking formulation for video

Includes an in-depth discussion of information processing in the cognitive process, embracing a new scheme of association memory and a new architecture for an artificial intelligent system with attractors of chaos

This complete guide to developing intelligent visual information processing systems is rich in examples, and will provide researchers and graduate students in computer vision and pattern recognition with a self-contained, invaluable and useful resource on the topic.

Contents:

Pattern Analysis and Statistical Learning 1

1.1 Introduction 1

1.1.1 Statistical Pattern Recognition 2

1.1.2 Pattern Theory 4

1.2 Statistical Classification 6

1.2.1 Feature Extraction and Selection 6

1.2.2 Classifier 7

1.3 Visual Pattern Representation 8

1.3.1 The Curse of Dimensionality 9

1.3.2 Dimensionality Reduction Techniques 9

1.4 Statistical Learning 10

1.4.1 Prediction Risk 11

1.4.2 Supervised, Unsupervised, and Others 12

1.5 Summary 14

References 14

2 Unsupervised Learning for Visual Pattern Analysis 15

2.1 Introduction 1

2.1.1 Unsupervised Learning I5

2.1.2 Visual Pattern Analysis 16

2.1.3 Outline 17

2.2 Cluster Analysis 17

2.3 Clustering Algorithms 21

2.3.1 Partitional Clustering 21

2.3.2 Hierarchical Clustering 30

2.4 Perceptual Grouping 33

2.4.1 Hierarchical Perceptual Grouping 33

2.4.2 Gestalt Grouping Principles 35

2.4.3 Contour Grouping 39

2.4.4 Region Grouping 45

2.5 Learning Representational Models for Visual Patterns 47

2.6 Summary 48

Appendix 48

References 48

3 Component Analysis 51

3.1 Introduction 51

3.2 Overview of Component Analysis 54

3.3 Generative Models 55

3.3.1 Principal Component Analysis 55

3.3.2 Nonnegative Matrix Factorization 66

3.3.3 Independent Component Analysis 72

3.4 Discriminative Models 16

3.4.1 Linear Discriminative Analysis 76

3.4.2 Oriented Component Analysis 79

3.4.3 Canonical Correlation Analysis 79

3.4.4 Relevant Component Analysis 81

3.5 Standard Extensions of the Linear Model 83

3.5.1 Latent Variable Analysis 83

3.5.2 Kernel Method 83

3.6 Summary 83

References 84

4 Manifold Learning 87

4.1 Introduction 87

4.2 Mathematical Preliminaries 91

4.2.1 Manifold Related Terminologies 91

4.2.2 Graph Related Terminologies 92

4.3 Global Methods 94

4.3.1 Multidimensional Scaling 94

4.3.2 Isometric Feature Mapping 95

4.3.3 Variants of the Isomap 96

4.4 Local Methods 100

4.4.1 Locally Linear Embedding 100

4.4.2 Laplacian Eigenmaps 103

4.4.3 Hessian Eigenmaps 107

4.4.4 Diffusion Maps 109

4.5 Hybrid Methods: Global Alignment of Local Models 113

4.5.1 Global Coordination of Local Linear Models 113

4.5.2 Charting a Manifold 115

4.5.3 Local Tangent Space Alignment 117

4.6 Summary 117

Appendix 118

References 118

5 Functional Approximation 121

5.1 Introduction 121

5.2 Modeling and Approximating the Visual Data 124

5.2.1 On Statistical Analysis 125

5.2.2 On Harmonic Analysis 126

5.2.3 Issues of Approximation and Compression 127

5.3 Wavelet Transform and Lifting Scheme 129

5.3.1 Wavelet Transform 129

5.3.2 Constructing a Wavelet Filter Bank 130

5.3.3 Lifting Scheme 132

5.3.4 Lifting-Based Integer Wavelet Transform 133

5.4 Optimal Integer Wavelet Transform 134

5.5 Introducing Adaptability into the Wavelet Transform 136

5.5.1 Curve Singularities in an Image 137

5.5.2 Anisotropic Basis 137

5.5.3 Adaptive Lifting-Based Wavelet 139

5.6 Adaptive Lifting Structure 140

5.6.1 Adaptive Prediction Filters 140

5.6.2 Adaptive Update Filters 142

5.7 Adaptive Directional Lifting Scheme 143

5.7.1 ADL Framework 144

5.7.2 Implementation of ADL 145

5.8 Motion Compensation Temporal Filtering in Video Coding 148

5.8.1 Overview of MCTF 148

5.8.2 MC in MCTF 151

5.8.3 Adaptive Lifting-Based Wavelets in MCTF 152

5.9 Summary and Discussions 153

References 154

6 Supervised Learning for Visual Pattern Classification 159

6.1 Introduction 159

6.2 An Example of Supervised Learning 160

6.3 Support Vector Machine 163

6.3.1 Optimal Separating Hyper-plane 163

6.3.2 Realization of SVM 167

6.3.3 Kernel Function 169

6.4 Boosting Algorithm 171

6.4.1 AdaBoost Algorithm 172

6.4.2 Theoretical Analysis of AdaBoost 173

6.4.3 AdaBoost Algorithm as an Additive Model 176

Summary 178

Appendix 178

References 179

7 Statistical Motion Analysis 181

7.1 Introduction 181

7.1.1 Problem Formulation 181

7.1.2 Overview of Computing Techniques 183

7.2 Bayesian Estimation of Optical Flow 186

7.2.1 Problem Formulation 186

7.2.2 MAP Estimation 190

7.2.3 Occlusion 192

7.3 Model-Based Motion Analysis 193

7.3.1 Motion Models 194

7.3.2 Statistical Model Selection 195

7.3.3 Learning Parameterized Models 196

7.4 Motion Segmentation 201

7.4.1 Layered Model: Multiple Motion Models 202

7.4.2 Clustering Optical Flow Field into Layers 204

7.4.3 Mixture Estimation for Layer Extraction 205

7.5 Statistics of Optical Flow 208

7.5.1 Statistics of Optical Flow 208

7.5.2 Motion Prior Modeling 210

7.5.3 Contrastive Divergence Learning 211

7.6 Summary 212

Appendix 212

References 214

8 Bayesian Tracking of Visual Objects 217

8.1 Introduction 217

8.2 Sequential Bayesian Estimation 219

8.2.1 Problem Formulation of Bayesian Tracking 220

8.2.2 Kalman Filter 221

8.2.3 Grid-Based Methods 222

8.2.4 Sub-optimal Filter 222

8.3 Monte Carlo Filtering 224

8.3.1 Problem Formulation 224

8.3.2 Sequential Importance Sampling 226

8.3.3 Sequential Monte Carlo Filtering 231

8.3.4 Particle Filter 232

8.4 Object Representation Model 235

8.4.1 Visual Learning for Object Representation 236

8.4.2 Active Contour 237

8.4.3 Appearance Model 241

8.5 Summary 243

References 244

9 Probabilistic Data Fusion for Robust Visual Tracking 245

9.1 Introduction 245

9.2 Earlier Work on Robust Visual Tracking 248

9.3 Data Fusion-Based Visual Tracker 251

9.3.1 Sequential Bayesian Estimator 251

9.3.2 The Four-Layer Data Fusion Visual Tracker 253

9.4 Layer 1: Visual Cue Fusion 255

9.4.1 Fusion Rules: Product Versus Weighted Sum 255

9.4.2 Adaptive Fusion Rule 257

9.4.3 Online Approach to Determining the Reliability of a Visual cue 258

9.5 Layer 2: Model Fusion 260

9.5.1 Pseudo-Measurement-Based Multiple Model Method 261

9.5.2 Likelihood Function 263

9.6 Layer 3: Tracker Fusion 264

9.6.1 Problem Formulation 265

9.6.2 Interactive Multiple Trackers 266

9.6.3 Practical Issues 267

9.7 Sensor Fusion 269

9.8 Implementation Issues and Empirical Results 271

9.8.1 Visual Cue Fusion Layer 271

9.8.2 Model Fusion Layer 274

9.8.3 Tracker Fusion Layer 276

9.8.4 Bottom-Up Fusion with a Three-Layer Structure 281

9.8.5 Multi-Censor Fusion Tracking System Validation 281

9.9 Summary 283

References 284

10 Multitarget Tracking in Video-Part I I287

10.1 Introduction 287

10.2 Overview of MTTV Methods 290

10.3 Static Model for Multitarget 292

10.3.1 Problem formulation 292

10.3.2 Observation Likelihood Function 294

10.3.3 Prior Model 295

10.4 Approximate Inference 296

10.4.1 Model Approximation 296

10.4.2 Algorithm Approximation 299

10.5 Fusing Information from Temporal and Bottom-Up Detectors 302

10.6 Experiments and Discussions 304

10.6.1 Proof-of-Concept 305

10.6.2 Comparison with Other Trackers 308

10.6.3 The Efficiency of the Gibbs Sampler 315

10.7 Summary 315

References 315

11 Multi-Target Tracking in Video - Part II 319

11.1 Introduction 319

11.2 Overview of the MTTV Data Association Mechanism 322

11.2.1 Handing Data Association Explicitly 322

11.2.2 Handing Data Association Implicitly 324

11.2.3 Detection and Tracking 325

11.3 The Generative Model for MTT 326

11.3.1 Problem Formulation 326

11.3.2 The Generative Model 327

11.4 Approximating The Marginal Term 329

11.4.1 The State Prediction 330

11.4.2 Existence and Association Posterior 332

11.5 Approximating the Interactive Term 334

11.6 Hybrid Measurement Process 335

11.7 Experiments and Discussion 335

11.7.1 Tracking Soccer Players 336

11.7.2 Tracking

Pedestrians in a Dynamic Scene 337

11.7.3 Discussion 337

11.8 Summary 340

References 340

12 Information Processing in Cognition Process and New Artificial Intelligent Systems 343

12.1 Introduction 343

12.2 Cognitive Model: A Prototype of Intelligent System 345

12.3 Issues in Theories and Methodologies of Current Brain Research and Vision Science 347

12.4 Interactive Behaviors and Selective Attention in the Process of Visual Cognition 351

12.5 Intelligent Information Processing and Modeling Based on Cognitive Mechanisms 353

12.5.1 Cognitive Modeling and Behavioral Control in Complex Systems in an Information Environment 353

12.5.2 Distributed Cognition 356

12.5.3 Neurophysiological Mechanism of Learning and Memory and Information Processing Model 358

12.6 Cognitive Neurosciences and Computational Neuroscience 359

12.6.1 Consciousness and Intention Reading 360

12.6.2 The Core of Computational Neuroscience is to Compute and Interpret the States of Nervous System 360

12.7 Soft Computing Method 360

12.8 Summary 361

References 362.

Notes:

Includes bibliographical references and index.

ISBN:

9781848823112

1848823118

9781848823129

1848823126

OCLC:

390942010