���� ��̥����ͼ�����

Contents
Preface
List of Abbreviations
Chapter 1 Introduction 1
1.1 Background 1
1.2 Contribution of This Book 2
1.3 Organization of This Book 3
References 4
Chapter 2 A Survey of Image Retrieval Techniques for Tire Pattern Database 6
2.1 Introduction 6
2.2 Tire Pattern Database and Performance Evaluation Methods 7
2.2.1 Tire pattern image databases 8
2.2.2 Performance evaluation 12
2.3 Tire Pattern Retrieval 16
2.3.1 Tire tread pattern retrieval 16
2.3.2 Tire surface wear feature extraction 21
2.3.3 Video tire pattern retrieval 23
2.3.4 Tire indentation mark image retrieval 24
2.3.5 Summary 27
2.4 Discussion about Future Research Directions 29
2.4.1 Standard test dataset and performance evaluation 29
2.4.2 Matching between tire indentation mark and tire tread pattern 30
2.5 Conclusions 30
References 31
Chapter 3 A Modified Tamura Feature for Tire Pattern Image Description 39
3.1 Introduction of Tamura Feature 39
3.2 Modification of Tamura Texture Feature 40
3.2.1 Tamura texture feature 40
3.2.2 Modification 44
3.3 Experimental Results 47
3.4 Conclusions 49
References 49
Chapter 4 H-SIFT: SIFT from High-Frequency Information of Tire Pattern Images 51
4.1 Introduction of SIFT Feature 51
4.2 Review of SIFT Feature 52
4.2.1 Scale space and relevant concepts 53
4.2.2 The model of Gaussian pyramid and difference of Gaussian pyramid 55
4.2.3 The establishment of the key points 57
4.2.4 The key points matching 59
4.3 Description of the Proposed Method H-SIFT 60
4.4 Experimental Results 62
4.5 Conclusions 64
References 64
Chapter 5 Study on Rotation-Invariant Texture Feature Extraction for Tire Pattern Retrieval 66
5.1 Introduction 66
5.2 Radon-DTCWT Algorithm 68
5.2.1 Radon transform 68
5.2.2 Translation sensitivity of ridgelet transform 69
5.2.3 The new Radon-DTCWT algorithm 72
5.3 Curvelet Energy Distribution Algorithm 75
5.3.1 Curvelet transform of tire pattern image 75
5.3.2 Direction characteristics of tire pattern images 76
5.3.3 Implementation of curvelet energy distribution algorithm 78
5.4 Experiment Results 80
5.5 Conclusions 83
References 83
Chapter 6 HOG-TT: A Robust HOG-Based Texture Feature Extraction Method Making Use of Texture Tendency in Tread Pattern Images 86
6.1 Introduction 86
6.2 Description of HOG-TT 88
6.2.1 HOG descriptor 88
6.2.2 HOG-TT 89
6.3 Experimental Results 93
6.4 Conclusions 97
References 97
Chapter 7 FF-TL: An E.ective Tread Pattern Image Classi��cation Algorithm Based on Transfer Learning 99
7.1 Introduction 99
7.2 Related Work 101
7.2.1 Convolutional neural network 101
7.2.2 Transfer learning 102
7.3 Proposed Algorithm 103
7.3.1 Fine-tuning the network 104
7.3.2 Feature extraction�� feature fusion and SVM classification 104
7.4 Experimental Results 105
7.4.1 Experimental dataset and performance evaluation parameter 105
7.4.2 Experimental results and analysis 106
7.5 Conclusions 108
References 109
Chapter 8 Summary and Future Work 113
8.1 Summary of the Book 113
8.2 Discussion of Future Work 115
8.3 Acknowledgment 116
Appendix 1: CIIP Tread Indentation Database 117
Appendix 2: CIIP Tread Pattern Database 118