Year: 2010
Type: dissertation
Abstract: viii ACKNOLWEDGEMENTS x ABBREVIATIONS xi CHAPTER 1 INTRODUCTION 1 1.1 Motivation 2 1.1.1 Representing Visual Content and Classifying Images 3 1.1.2 Ambiguity of Natural Language in Text Captions 4 1.1.3 Use of Hybrid Visual and Textual Metadata Models 5 1.2 Research Objectives 6 1.3 Structure of this Thesis 8 CHAPTER 2 FUNDAMENTALS 10 2.1 Main Processes for IMR 10 2.2 Content-Based Image Retrieval (CBIR) 13 2.2.1 Global Features 14 2.2.2 Local Features 15 2.3 Semantic-Based Image Retrieval (SBIR) 16 2.3.1 Knowledge (Ontology-based) Representation Techniques 18 2.3.2 Advantages of Using Ontologies for IR 21 2.4 MPEG-7 and Ontology-based KB 21 2.5 Summary 25 CHAPTER 3 SURVEY AND ANALYSIS OF THE STATE-OF-THE-ART FRAMEWORKS 26 3.1 Problem Analysis of the Image Retrieval Systems 26 3.2 Formal Requirements of Image Retrieval Systems 28 3.3 Survey of State of the Art Frameworks 29 3.3.1 Ontology-Based KB Frameworks for IMR 29 3.3.2 Visual Features-Based Frameworks for Visual Content Representation 34 3.4 Discussion 39 3.5 Summary 42
Author: Kraisak Kesorn
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W788255342