Design, implementation, and evaluation of scalable content-based image retrieval techniques.

Wong, Yuk Man.Thesis (M.Phil.)--Chinese University of Hong Kong, 2007.Includes bibliographical references (leaves 95-100).Abstracts in English and Chinese.Abstract --- p.iiAcknowledgement --- p.vChapter 1 --- Introduction --- p.1Chapter 1.1 --- Overview --- p.1Chapter 1.2 --- Contribution --- p.3Chapter 1.3 --- Organization of This Work --- p.5Chapter 2 --- Literature Review --- p.6Chapter 2.1 --- Content-based Image Retrieval --- p.6Chapter 2.1.1 --- Query Technique --- p.6Chapter 2.1.2 --- Relevance Feedback --- p.7Chapter 2.1.3 --- Previously Proposed CBIR systems --- p.7Chapter 2.2 --- Invariant Local Feature --- p.8Chapter 2.3 --- Invariant Local Feature Detector --- p.9Chapter 2.3.1 --- Harris Corner Detector --- p.9Chapter 2.3.2 --- DOG Extrema Detector --- p.10Chapter 2.3.3 --- Harris-Laplacian Corner Detector --- p.13Chapter 2.3.4 --- Harris-Affine Covariant Detector --- p.14Chapter 2.4 --- Invariant Local Feature Descriptor --- p.15Chapter 2.4.1 --- Scale Invariant Feature Transform (SIFT) --- p.15Chapter 2.4.2 --- Shape Context --- p.17Chapter 2.4.3 --- PCA-SIFT --- p.18Chapter 2.4.4 --- Gradient Location and Orientation Histogram (GLOH) --- p.19Chapter 2.4.5 --- Geodesic-Intensity Histogram (GIH) --- p.19Chapter 2.4.6 --- Experiment --- p.21Chapter 2.5 --- Feature Matching --- p.27Chapter 2.5.1 --- Matching Criteria --- p.27Chapter 2.5.2 --- Distance Measures --- p.28Chapter 2.5.3 --- Searching Techniques --- p.29Chapter 3 --- A Distributed Scheme for Large-Scale CBIR --- p.31Chapter 3.1 --- Overview --- p.31Chapter 3.2 --- Related Work --- p.33Chapter 3.3 --- Scalable Content-Based Image Retrieval Scheme --- p.34Chapter 3.3.1 --- Overview of Our Solution --- p.34Chapter 3.3.2 --- Locality-Sensitive Hashing --- p.34Chapter 3.3.3 --- Scalable Indexing Solutions --- p.35Chapter 3.3.4 --- Disk-Based Multi-Partition Indexing --- p.36Chapter 3.3.5 --- Parallel Multi-Partition Indexing --- p.37Chapter 3.4 --- Feature Representation --- p.43Chapter 3.5 --- Empirical Evaluation --- p.44Chapter 3.5.1 --- Experimental Testbed --- p.44Chapter 3.5.2 --- Performance Evaluation Metrics --- p.44Chapter 3.5.3 --- Experimental Setup --- p.45Chapter 3.5.4 --- Experiment I: Disk-Based Multi-Partition Indexing Approach --- p.45Chapter 3.5.5 --- Experiment II: Parallel-Based Multi-Partition Indexing Approach --- p.48Chapter 3.6 --- Application to WWW Image Retrieval --- p.55Chapter 3.7 --- Summary --- p.55Chapter 4 --- Image Retrieval System for IND Detection --- p.60Chapter 4.1 --- Overview --- p.60Chapter 4.1.1 --- Motivation --- p.60Chapter 4.1.2 --- Related Work --- p.61Chapter 4.1.3 --- Objective --- p.62Chapter 4.1.4 --- Contribution --- p.63Chapter 4.2 --- Database Construction --- p.63Chapter 4.2.1 --- Image Representations --- p.63Chapter 4.2.2 --- Index Construction --- p.64Chapter 4.2.3 --- Keypoint and Image Lookup Tables --- p.67Chapter 4.3 --- Database Query --- p.67Chapter 4.3.1 --- Matching Strategies --- p.68Chapter 4.3.2 --- Verification Processes --- p.71Chapter 4.3.3 --- Image Voting --- p.75Chapter 4.4 --- Performance Evaluation --- p.76Chapter 4.4.1 --- Evaluation Metrics --- p.76Chapter 4.4.2 --- Results --- p.77Chapter 4.4.3 --- Summary --- p.81Chapter 5 --- Shape-SIFT Feature Descriptor --- p.82Chapter 5.1 --- Overview --- p.82Chapter 5.2 --- Related Work --- p.83Chapter 5.3 --- SHAPE-SIFT Descriptors --- p.84Chapter 5.3.1 --- Orientation assignment --- p.84Chapter 5.3.2 --- Canonical orientation determination --- p.84Chapter 5.3.3 --- Keypoint descriptor --- p.87Chapter 5.4 --- Performance Evaluation --- p.88Chapter 5.5 --- Summary --- p.90Chapter 6 --- Conclusions and Future Work --- p.92Chapter 6.1 --- Conclusions --- p.92Chapter 6.2 --- Future Work --- p.93Chapter A --- Publication --- p.94Bibliography --- p.9