Machine Learning and Statistical Modeling Approaches to Image Retrieval (eBook)

Contents 7
Preface 13
Acknowledgments 16
Chapter 1 INTRODUCTION 17
1. Text-Based Image Retrieval 17
2. Content-Based Image Retrieval 19
3. Automatic Linguistic Indexing of Images 20
4. Applications of Image Indexing and Retrieval 21
4.1 Web- Related Applications 21
4.2 Biomedical Applications 23
4.3 Space Science 24
4.4 Other Applications 26
5. Contributions of the Book 26
5.1 A Robust Image Similarity Measure 26
5.2 Clustering- Based Retrieval 27
5.3 Learning and Reasoning with Regions 28
5.4 Automatic Linguistic Indexing 28
5.5 Modeling Ancient Paintings 29
6. The Structure of the Book 30
Chapter 2 IMAGE RETRIEVAL AND LINGUISTIC INDEXING 31
1. Introduction 31
2. Content-Based Image Retrieval 31
2.1 Similarity Comparison 32
2.2 Semantic Gap 34
3. Categorization and Linguistic Indexing 36
4. Summary 39
Chapter 3 MACHINE LEARNING AND STATISTICAL MODELING 41
1. Introduction 41
2. Spectral Graph Clustering 41
3. VC Theory and Support Vector Machines 44
3.1 VC Theory 45
3.2 Support Vector Machines 46
4. Additive Fuzzy Systems 50
5. Support Vector Learning for Fuzzy Rule-Based Classification Systems 52
5.1 Additive Fuzzy Rule- Based Classification Systems 53
5.2 Positive Definite Fuzzy Classifiers 54
5.3 An SVM Approach to Build Positive Definite Fuzzy Classifiers 56
6. 2-D Multi-Resolution Hidden Markov Models 58
7. Summary 62
Chapter 4 A ROBUST REGION-BASEDSIMILARITY MEASURE 63
1. Introduction 63
2. Image Segmentation and Representation 65
2.1 Image Segmentation 65
2.2 Fuzzy Feature Representation of an Image 67
2.3 An Algorithmic View 71
3. Unified Feature Matching 72
3.1 Similarity Between Regions 72
3.2 Fuzzy Feature Matching 74
3.3 The UFM Measure 76
3.4 An Algorithmic View 78
4. An Algorithmic Summarization of the System 79
5. Experiments 80
5.1 Query Examples 80
5.2 Systematic Evaluation 80
5.3 Speed 87
5.4 Comparison of Membership Functions 88
6. Summary 89
Chapter 5 CLUSTER-BASED RETRIEVALBY UNSUPERVISED LEARNING 91
1. Introduction 91
2. Retrieval of Similarity Induced Image Clusters 92
2.1 System Overview 92
2.2 Neighboring Target Images Selection 93
2.3 Spectral Graph Partitioning 94
2.4 Finding a Representative Image for a Cluster 95
3. An Algorithmic View 96
3.1 Outline of Algorithm 96
3.2 Organization of Clusters 98
3.3 Computational Complexity 99
3.4 Parameters Selection 100
4. A Content-Based Image Clusters Retrieval System 101
5. Exper iments 102
5.1 Query Examples 103
5.2 Systematic Evaluation 103
5.3 Speed 109
5.4 Application of CLUE to Web Image Retrieval 110
6. Summary 114
Chapter 6 CATEGORIZATION BY LEARNING AND REASONING WITH REGIONS 115
1. Introduction 115
2. Learning Region Prototypes Using Diverse Density 118
2.1 Diverse Density 118
2.2 Learning Region Prototypes 120
2.3 An Algorithmic View 121
3. Categorization by Reasoning with Region Prototypes 122
3.1 A Rule- Based Image Classifier 122
3.2 Support Vector Machine Concept Learning 124
3.3 An Algorithmic View 126
4. Experiments 126
4.1 Experiment Setup 127
4.2 Categorization Results 129
4.3 Sensitivity to Image Segmentation 131
4.4 Sensitivity to the Number of Categories 131
4.5 Sensitivity to the Size and Diversity of Training Set 134
4.6 Speed 136
5. Summary 136
Chapter 7 AUTOMATIC LINGUISTIC INDEXING OF PICTURES 139
1. Introduction 139
2. System Architecture 141
2.1 Feature Extraction 141
2.2 Multiresolution Statistical Modeling 142
2.3 Statistical Linguistic Indexing 144
2.4 Major Advantages 144
3. Model-Based Learning of Concepts 144
4. Automatic Linguistic Indexing of Pictures 146
5. Experiments 148
5.1 Training Concepts 148
5.2 Performance with a Controlled Database 149
5.3 Categorization and Annotation Results 152
6. Summary 154
Chapter 8 MODELING ANCIENT PAINTINGS 157
1. Introduction 157
2. Mixture of 2-D Multi-Resolution Hidden Markov Models 160
3. Feature Extraction 161
4. System Architecture 164
5. Experiments 166
5.1 Background on the Artists 166
5.2 Extract Stroke/Wash Styles by the Mixture Model 167
5.3 Classification Results 171
6. Other Applications 174
7. Summary 177
Chapter 9 CONCLUSIONS AND FUTURE WORK 179
1. Summary 179
1.1 A Robust Region- Based Image Similarity Measure 179
1.2 Cluster- Based Retrieval of Images by Unsupervised Learning 181
1.3 Image Categorization by Learning and Reasoning with Regions 182
1.4 Automatic Linguistic Indexing of Pictures 183
1.5 Characterization of Fine Art Painting Styles 184
2. Future Work 185
References 189
Index 197
More eBooks at www.ciando.com 0

2.1 Similarity Comparison (p.16-17)

Similarity comparison is a key issue in CBIR [Santini and Jain, 1999]. In general, the comparison is performed over imagery features. According to the scope of representation, features fall roughly into two categories: global features and local features. The former category includes texture histogram, color histogram, color layout of the whole image, and features selected from multidimensional discriminant analysis of a collection of images [Faloutsos et al., 1994; Gupta and Jain, 1997; Pentland et al., 1996; Smith and Chang, 1996; Swets and Weng, 1996]. In the latter category are color, texture, and shape features for subimages [Picard and Minka, 1995], segmented regions [Carson et al., 2002; Chen and Wang, 2002; Ma and Manjunath, 1997; Wang et al., 2001b], and interest points [Schmid and Mohr, 1997].

As a relatively mature method, histogram matching has been applied to many general-purpose image retrieval systems such as IBM QBIC [Faloutsos et al., 1994], MIT Photobook [Pentland et al., 1996], Virage System [Gupta and Jain, 1997], and Columbia VisualSEEK and WebSEEK [Smith and Chang, 1996], etc. The Mahalanobis distance [Hafner et al., 1995] and intersection distance [Swain and Ballard, 1991] are commonly used to compute the difference between two histograms with the same number of bins. When the number of bins are different, e.g., when a sparse representation is used, the Earth Mover’s Distance (EMD) [Rubner et al., 1997] applies. The EMD is computed by solving a linear programming problem. A major drawback of the global histogram search lies in its sensitivity to intensity variations, color distortions, and cropping.

Many approaches have been proposed to tackle this problem:

* The PicToSeek [Gevers and Smeulders, 2000] system uses color models invariant to object geometry, object pose, and illumination.

* VisualSEEK and Virage systems attempt to reduce the influence of intensity variations and color distortions by employing spatial rela tionships and color layout in addition to those elementary color, texture, and shape features.

* The same idea of color layout indexing is extended in a later system, Stanford WBIIS [Wang et al., 1998], which, instead of averaging, characterizes the color variations over the spatial extent of an image by Daubechies’ wavelet coefficients and their variances.

* Schmid and Mohr [Schmid and Mohr, 1997] proposed a method of indexing images based on local features of automatically detected interest points of images.

* Minka and Picard [Minka and Picard, 1997] described a learning algorithm for selecting and grouping features. The user guides the learning process by providing positive and negative examples.

* The approach presented in [Swets and Weng, 1996] uses what is called the Most Discriminating Features for image retrieval. These features are extracted from a set of training images by optimal linear projection.

* The Virage system allows users to adjust weights of implemented features according to their own perceptions. The PicHunter system [Cox et al., 2000] and the UIUC MARS [Mehrotra et al., 1997] system are self-adaptable to different applications and different users based upon user feedbacks.

* To approximate the human perception of the shapes of the objects in the images, Del Bimbo and Pala [Bimbo and Pala, 1997] introduced a measure of shape similarity using elastic matching.

* In [Mojsilovic et al., 2000], matching and retrieval are performed along what is referred to as perceptual dimensions which are obtained from subjective experiments and multidimensional scaling based on the model of human perception of color patterns.

* In [Berretti et al., 2000], two distinct similarity measures, concerning respectively with fitting human perception and with the efficiency of data organization and indexing, are proposed for content-based image retrieval by shape similarity.