Hybrid Soft Computing for Image Segmentation (eBook)

Foreword 6
Preface 9
Contents 13
Hybrid Swarms Optimization Based Image Segmentation 15
1 Introduction 15
2 Preliminaries 18
2.1 Firefly Algorithm (FA) 18
2.2 Social Spider Optimization Algorithm (SSO) 19
3 The Proposed Hybrid Segmentation Algorithm 22
4 Experiments and Discussion 22
5 Conclusion and Future Work 33
References 33
Grayscale Image Segmentation Using Multilevel Thresholding and Nature-Inspired Algorithms 36
1 Introduction 36
2 Formulation of Multilevel Thresholding Problem 38
2.1 Kapur's Entropy Criterion 38
2.2 Otsu Criterion 39
3 Nature-Inspired Algorithms Based Multilevel Thresholding 40
3.1 Swarm Intelligence-Based Optimization Algorithms Based Multilevel Thresholding 40
3.2 Evolutionary Algorithms Based Multilevel Thresholding 41
3.3 Physics-Based Algorithms Based Multilevel Thresholding 41
3.4 Hybrid NAs Based Multilevel Thresholding 42
4 GSA-GA Based Multilevel Image Thresholding 42
4.1 GSA-GA 44
4.2 Implementation of GSA-GA for Multilevel Thresholding 45
5 Experimental Results and Discussion 45
5.1 Test Images 46
5.2 Experimental Settings 46
5.3 Performance Metrics 47
5.4 Experiment 1: Maximizing Kapur's Entropy 48
5.5 Experiment 2: Maximizing Otsu 48
5.6 Running Time Analysis Using Student's t-test 57
6 Conclusion 62
References 62
A Novel Hybrid CS-BFO Algorithm for Optimal Multilevel Image Thresholding Using Edge Magnitude Information 66
1 Introduction 67
2 Edge Magnitude Computation 69
3 Evolutionary Computing Techniques 72
3.1 Cuckoo Search Technique 72
3.2 Genetic Algorithm 72
3.3 Particle Swarm Optimization 73
3.4 Hybrid CS-BFO Algorithm 74
4 Proposed Methodology 77
5 Results and Discussions 79
6 Conclusion 96
References 96
REFII Model and Fuzzy Logic as a Tool for Image Classification Based on Image Example 99
1 Introduction 100
2 Background 100
3 REFII Model and Introduction 102
4 Conceptual Solution Proposal 104
5 Realization of Proposed Solution 106
6 Model Application 116
7 Future Research 117
8 Conclusion 118
References 119
Microscopic Image Segmentation Using Hybrid Technique for Dengue Prediction 121
1 Introduction 122
2 Dengue Scenario All over the World 123
3 The Dengue Virus and Its Effect in Human Body 124
4 Dengue Diagnosis and Treatment 128
4.1 Dengue Diagnosis 128
4.2 Dengue Treatment 129
5 Platelet Segmentation 131
5.1 Image Acquisition 131
5.2 Platelet Segmentation Using Color-Based Segmentation 132
5.3 Platelet Segmentation Using K-Means Clustering Technique 132
5.4 Platelet Segmentation Using Hybrid Soft Computing Technique 133
6 Conclusion 145
References 145
Extraction of Knowledge Rules for the Retrieval of Mesoscale Oceanic Structures in Ocean Satellite Images 149
1 Introduction 149
1.1 Contributions of the Chapter 151
1.2 Related Work 152
1.3 Structure of the Chapter 153
2 Dataset 153
3 Workflow of the Tool 154
3.1 Extraction of Knowledge Rules 157
4 Experimental Results 159
4.1 Knowledge Rules 159
4.2 Labeling 161
5 Implementation 164
5.1 User Interface 164
5.2 Class Modeling 168
5.3 Data Modeling 168
6 Conclusions and Future Work 171
References 172
Hybrid Uncertainty-Based Techniques for Segmentation of Satellite Imagery and Applications 175
1 Introduction 176
1.1 Edge-Based Image Segmentation 177
1.2 Threshold-Based Image Segmentation [7] 177
1.3 Region-Based Image Segmentation [7] 178
1.4 PDE-Based Image Segmentation [7] 178
1.5 ANN-Based Image Segmentation [7] 178
1.6 Uncertainty-Based Image Segmentation [7] 178
2 Background 178
3 Uncertainty-Based Techniques 181
3.1 Fuzzy C-Means 182
3.2 Rough C-Means 182
3.3 Intuitionistic Fuzzy C-Means 184
4 Hybrid Techniques 184
4.1 The Hybrid Algorithms 185
5 Case Study 189
5.1 Discussion on the Results 191
6 Scope for Future Work 192
7 Conclusions 193
References 194
Improved Human Skin Segmentation Using Fuzzy Fusion Based on Optimized Thresholds by Genetic Algorithms 196
1 Introduction 197
2 Background 198
3 Improved Human Skin Segmentation 200
3.1 Seed Detection 203
3.2 Edge Detection 204
3.3 Seed Expansion and Local Probability Map 205
3.4 Fuzzy Fusion 205
3.5 Genetic Algorithms 207
4 Experimental Results 209
4.1 Data Sets 210
4.2 Evaluation Settings 211
4.3 Statistical Analysis 212
4.4 Qualitative Analysis 214
5 Conclusions 215
References 216
Uncertainty-Based Spatial Data Clustering Algorithms for Image Segmentation 219
1 Introduction 219
2 Based Models 220
2.1 Fuzzy Sets 221
2.2 Rough Sets 221
2.3 Hybrid Models 221
3 Uncertainty-Based Data Clustering Algorithms 221
3.1 Fuzzy C-Means Algorithm (FCM) 222
3.2 RFCM 222
3.3 IFCM 223
3.4 RIFCM [5] 223
4 Image Segmentation 224
5 Spatial Data Clustering Algorithms in Image Segmentation 224
5.1 Spatial Information 225
6 Conclusion 235
References 236
Coronary Artery Segmentation and Width Estimation Using Gabor Filters and Evolutionary Computation Techniques 238
1 Introduction 239
2 Background 240
2.1 Single-Scale Gabor Filters (SSG) 240
2.2 Evolutionary Computation Techniques 242
3 Proposed Method 249
3.1 Optimal Parameter Selection of SSG 250
3.2 Thresholding of the SSG Filter Response 251
3.3 Postprocessing of Segmented Vessels 252
4 Computational Experiments 254
4.1 Results of Coronary Artery Detection 254
4.2 Results of Coronary Artery Segmentation 256
4.3 Vessel Width Estimation 257
5 Concluding Remarks 260
References 261
Hybrid Intelligent Techniques for Segmentation of Breast Thermograms 263
1 Introduction 264
2 Breast Cancer 267
2.1 Different Risk Factors of Breast Cancer 269
2.2 Importance of Breast Cancer Awareness 272
3 The Role of Thermal Breast Imaging in the Identification of Breast Cancer 273
3.1 Predictive Ability of Breast Thermography 273
3.2 Interpretation of Breast Thermography 274
4 Breast Thermal Image Acquisition 276
4.1 Laboratory and Patient Preparation 276
4.2 Acquisition System 277
4.3 Capturing Views and Number of Captured Images 278
4.4 Available Databases for Research 281
5 Region of Interest Segmentation from the Thermal Breast Image 282
5.1 Breast Region Segmentation 283
5.2 Hottest Region Segmentation 285
6 Conclusion 293
References 293
Modeling of High-Dimensional Data for Applications of Image Segmentation in Image Retrieval and Recognition Tasks 298
1 Introduction 298
2 Hybrid Multidimensional Modeling of Feature Vectors 299
2.1 N-Dimensional Probability Distribution Functions in PFC Modeling 300
3 Discussion in Details Over PFC Approach 304
3.1 Interpolation of Some Complicated Functions Using Combinations of a Simple Function 304
3.2 Only Local Changes of the Curve if One Node Is Exchanged 305
3.3 No Matter if the Curve Is Opened or Closed 305
3.4 Data Extrapolation Is Computed via the Same Formulas As Interpolation 305
3.5 Object Modeling in any Dimension N 306
3.6 Curve Parameterization 306
3.7 Modeling of Specific and Nontypical Curves: Signatures, Fonts, Symbols, Characters, or Handwriting 306
3.8 Reconstruction of Irregular Shapes 306
3.9 Applications in Numerical Analysis Because of Very Precise Interpolation of Unknown Values 306
3.10 Even for only Two Nodes a Curve Can Be Modeled 306
4 Image Retrieval via PFC High-Dimensional Feature Reconstruction 307
4.1 Grey Scale Image Retrieval Using PFC 3D Method 307
4.2 Color Image Retrieval via PFC Method 308
5 Recognition Tasks via High-Dimensional Feature Vectors' Interpolation 311
5.1 Signature Modeling and Multidimensional Recognition 311
5.2 Modeling and Interpolation of Nontypical Curves and Irregular Shapes 316
6 Result Analysis 321
7 Conclusions 323
References 323
Index 325