Adaptive Image Processing A Computational Intelligence Perspective

1.1 The Importance of Vision......Page 1
Table of Contents......Page 255
References......Page 259
1.2 Adaptive Image Processing......Page 2
Edges......Page 3
Textures......Page 4
1.4 Difficulties in Adaptive Image Processing System Design......Page 5
Optimization......Page 7
1.5 Computational Intelligence Techniques......Page 8
Neural Networks......Page 10
Fuzzy Logic......Page 11
Evolutionary Computation......Page 12
1.6 Scope of the Book......Page 13
1.6.1 Image Restoration......Page 14
1.6.2 Edge Characterization and Detection......Page 18
1.7.1 Application of Neural Networks for Image Restoration......Page 19
1.7.3 Application of Fuzzy Set Theory to Adaptive Regularization......Page 20
1.8 Overview of This Book......Page 21
2.1 Image Distortions......Page 23
2.2.1 Degradation Measure......Page 26
2.2.2 Neural Network Restoration......Page 27
2.3 Neural Network Restoration Algorithms in the Literature......Page 29
2.4 An Improved Algorithm......Page 32
2.5 Analysis......Page 34
2.6 Implementation Considerations......Page 36
2.7.2 Efficiency......Page 37
2.7.3 An Application Example......Page 38
2.8 Summary......Page 39
3.1 Introduction......Page 41
3.2 Dealing with Spatially Variant Distortion......Page 43
3.3.1 Motivation......Page 46
3.3.2 The Gradient-Based Method......Page 48
3.3.3 Local Statistics Analysis......Page 57
3.4 Correcting Spatially Variant Distortion Using Adaptive Constraints......Page 66
3.5 Semi-Blind Restoration Using Adaptive Constraints......Page 68
3.6 Implementation Considerations......Page 72
3.7.1 Efficiency......Page 73
3.7.2 An Application Example......Page 74
3.8.1 Problem Formulation......Page 76
3.8.2 Problem Solution......Page 77
3.8.3 Conditions for KKT Theory to Hold......Page 80
3.9 Summary......Page 82
4.1 Introduction......Page 84
4.2 Motivation......Page 85
4.3 A LVMSE-Based Cost Function......Page 86
4.3.1 The Extended Algorithm for the LVMSE-Modified Cost Function......Page 87
4.3.2 Analysis......Page 91
4.4 A Log LVMSE-Based Cost Function......Page 95
4.4.1 The Extended Algorithm for the Log LVR-Modified Cost Function......Page 96
4.4.2 Analysis......Page 98
4.6.1 Color Image Restoration......Page 101
4.6.4 Robustness Evaluation......Page 104
4.7 Summary......Page 108
5.1 Model-Based Neural Network......Page 109
5.1.1 Weight-Parameterized Model-Based Neuron......Page 110
5.2 Hierarchical Neural Network Architecture......Page 111
5.4 HMBNN for Adaptive Image Processing......Page 113
5.6 Adaptive Regularization: An Alternative Formulation......Page 114
5.6.1 Correspondence with the General HMBNN Architecture......Page 116
5.7 Regional Training Set Definition......Page 120
5.8 Determination of the Image Partition......Page 122
5.9 The Edge-Texture Characterization......Page 124
5.10 The ETC Fuzzy HMBNN for Adaptive Regularization......Page 128
5.11 Theory of Fuzzy Sets......Page 129
5.12 Edge-Texture Fuzzy Model Based on ETC Measure......Page 131
5.13 Architecture of the Fuzzy HMBNN......Page 133
5.14 Estimation of the Desired Network Output......Page 135
5.15 Fuzzy Prediction of Desired Gray Level Value......Page 136
5.15.1 Definition of the Fuzzy Estimator Membership Function......Page 137
5.15.2 Fuzzy Inference Procedure for Predicted Gray Level Value......Page 138
5.15.3 Defuzzification of the Fuzzy Set G......Page 139
5.15.4 Regularization Parameter Update......Page 140
5.15.5 Update of the Estimator Fuzzy Set Width Parameters......Page 142
5.16 Experimental Results......Page 143
5.17 Summary......Page 152
6.1 Introduction......Page 153
6.2.1 Genetic Algorithm......Page 154
6.2.2 Evolutionary Strategy......Page 155
6.2.3 Evolutionary Programming......Page 156
6.3 The ETC-pdf Image Model......Page 158
6.4 Adaptive Regularization Using Evolutionary Programming......Page 162
6.4.1 Competition under Approximate Fitness Criterion......Page 166
6.4.2 Choice of Optimal Regularization Strategy......Page 167
6.5 Experimental Results......Page 170
6.6.1 Hierarchical Cluster Model......Page 177
6.7 Summary......Page 178
7.1 Introduction......Page 180
7.1.2 Blur Identification by Recursive Soft Decision......Page 182
7.2.1 Formulation of Blind Image Deconvolution as an Evolutionary Strategy......Page 183
7.2.2 Knowledge-Based Reinforced Mutation......Page 190
7.2.3 Perception-Based Image Restoration......Page 194
7.2.4 Recombination Based on Niche-Space Residency......Page 196
7.2.5 Performance Evaluation and Selection......Page 198
7.3.1 Recursive Subspace Optimization......Page 200
7.3.2 Hierarchical Neural Network for Image Restoration......Page 201
7.3.3 Soft Parametric Blur Estimator......Page 207
7.3.4 Blur Identification by Conjugate Gradient Optimization......Page 208
7.3.5 Blur Compensation......Page 211
7.4 Simulation Examples......Page 213
7.4.1 Identification of 2D Gaussian Blur......Page 214
7.4.2 Identification of 2D Gaussian Blur from Degraded Image with Additive Noise......Page 216
7.4.3 Identification of 2D Uniform Blur by CRL......Page 217
7.5 Conclusions......Page 222
8.1 Introduction......Page 225
8.2.1 Input-Parameterized Model-Based Neuron......Page 226
8.2.3 Edge Characterization and Detection......Page 228
8.3 Network Architecture......Page 230
8.3.3 Neuron in Sub-Network......Page 232
8.3.4 Dynamic Tracking Neuron......Page 233
8.3.5 Binary Edge Configuration......Page 234
8.3.6 Correspondence with the General HMBNN Architecture......Page 235
8.4.1 Determination of for Sub-Network......Page 236
8.5 Recognition Stage......Page 237
8.5.2 Identification of Secondary Edge Points......Page 238
8.6 Experimental Results......Page 239
8.7 Summary......Page 243
Adaptive Image Processing, A Computational Intelligence Perspective......Page 249
Preface......Page 253
Acknowledgments......Page 254