基于语义的图像检索

PrefaceList of AbbreviationsChapter 1 Introduction1．1 Background1．1．1 The 'Semantic Gap1．1．2 Query by Keywords1．2 Objectives1．3 Contributions of this Book1．3．1 Identifying Existing Semantic Learning Techniques1．3．2 Designing Effective Feature Extraction Methods for Arbitrary-Shaped Regions\"1．3．3 High-Level Concept Learning Using Decision Tree1．3．4 Applying RBIR with Semantics to Web Image Search1．4 Organization of the BookChapter 2 Key Techniques in Semantic-Based Image Retrieval2．1 Introduction2．2 Techniques and Issues in Region-Based Image Retrieval2．2．1 Image Segmentation2．2．2 Low-Level Image Feature Extraction2．2．3 Similarity Measure2．2．4 Test Database and Performance Evaluation2．3 High-Level Image Semantic Learning Techniques2．3．1 Object-Ontology2．3．2 Machine Learning2．3．3 Relevance Feedback (RF)2．3．4 Semantic Template2．3．5 Fusion of Multiple Resources for Web Image Search2．3．6 Deep Learning2．3．7 Summary of Existing Techniques in Image Semantic Learning2．4 Research Problems Addressed in this BookChapter 3 Deriving Image Semantics from Color Features3．1 Introduction3．2 Region Color Feature Extraction and Semantic Color Naming3．2．1 Region Color Features3．2．2 Semantic Color Names3．3 Image Retrieval using Semantic Color Names3．3．1 RBIR with Semantic Color Names3．3．2 Feature Normalization3．3．3 Image Similarity Measure using EMD3．4 Results and Analysis3．4．1 Test Database and Performance Evaluation Model3．4．2 Comparison of Different Color Features3．4．3 Performance of the Proposed Color Naming Method3．4．4 Image Retrieval with Color Names， Region Color Features and GlobalColor Features3．5 Discussion and ConclusionsChapter 4 Effective Texture Feature Extraction from Arbitrary-Shaped Regions4．1 Introduction4．2 Deriving Texture Features from Arbitrary-Shaped Regions4．2．1 Projection onto Convex Set (POCS) Theory4．2．2 Extracting Region Texture Features Using POCS-ER4．2．3 Theoretical Analysis of POCS-ER4．2．4 Implementation of POCS-ER4．3 POCS-ER on Brodatz Textures4．3．1 Illustration of POCS-ER Process4．3．2 Performance of POCS-ER Measured by PSNR4．3．3 Performance of POCS-ER Measured by Retrieval Performance4．4 POCS-ER for Real-World Image Retrieval4．4．1 Experimental Setups4．4．2 Performance of Different Texture Feature Extraction Methods in RBIR．．．4．4．3 RBIR with Color， Texture， Color & Texture4．4．4 Comparison of Region Features and Global Features in Image Retrieval4．5 Conclusions and DiscussionChapter 5 Deriving High-Level Image Concepts Using Decision Tree Learning5．1 Introduction5．2 Decision Tree Learning5．2．1 Overview5．2．2 Decision Tree Induction for Image Semantic Learning5．3 The Proposed Decision Tree Induction Algorithm DT-ST5．3．1 Semantic Template Construction5．3．2 Image Feature Discretization5．3．3 Decision Tree Induction5．4 Results and Analysis5．4．1 Selection of Pre-pnming Threshold5．4．2 Pruning Unknowns5．4．3 Handling Queries with Concepts outside the Training Concept Set5．4．4 Comparison of DT-ST with ID3 and C4．55．5 Region-Based Image Retrieval with High-Level Semantics5．6 Discussion5．6．1 Scalability of DT-ST5．6．2 The Advantage of Image Retrieval with High-Level Concepts5．7 ConclusionsChapter 6 Application of Semantic-Based RBIR to Web Image Search6．1 Introduction6．2 The False Filtering Algorithm6．3 Results and Analysis6．3．1 Web Image Collection and Performance Evaluation6．3．2 Experimental Results6．4 Discussions6．4．1 Integration6．4．2 FF Response Time6．4．3 Scalability6．5 ConclusionsChapter 7 Conclusions and Future Work7．1 Conclusions of this Book7．2 Future Research DirectionsBibliographyAppendix A HSV Color Histogram and HSV-RGB ConversionAppendix B Tamura Texture FeaturesAppendix C lllustration of POCS-ER Process Using ZR and MPAppendix D Pre-pruning &Post-pruning in DT-ST