Intelligent Image Databases: Towards Advanced Image Retrieval(421 The Springer International Series in Engineering and Computer Science)

Intelligent Image Databases: Towards Advanced Image Retrieval addresses the image feature selection issue in developing content-based image retrieval systems. The book first discusses the four important issues in developing a complete content-based image retrieval system, and then demonstrates that image feature selection has significant impact on the remaining issues of system design. Next, it presents an in-depth literature survey on typical image features explored by contemporary content-based image retrieval systems for image matching and retrieval purposes. The goal of the survey is to determine the characteristics and the effectiveness of individual features, so as to establish guidelines for future development of content-based image retrieval systems. Intelligent Image Databases: Towards Advanced Image Retrieval describes the Advanced Region-Based Image Retrieval System (ARBIRS) developed by the authors for color images of real-world scenes. They have selected image regions for building ARBIRS as the literature survey suggests that prominent image regions, along with their associated features, provide a higher probability for achieving a higher level content-based image retrieval system. A major challenge in building a region-based image retrieval system is that prominent regions are rather difficult to capture in an accurate and error-free condition, particularly those in images of real-world scenes. To meet this challenge, the book proposes an integrated approach to tackle the problem via feature capturing, feature indexing, and database query. Through comprehensive system evaluation, it is demonstrated how these systematically integrated efforts work effectively to accomplish advanced image retrieval. Intelligent Image Databases: Towards Advanced Image Retrieval serves as an excellent reference and may be used as a text for advanced courses on the topic.

Table of Contents:
1. Introduction.- 1.1 The Need for Intelligent Image Databases.- 1.2 Significance of Feature Space Selection.- 1.3 Towards Advanced Image Retrieval.- 1.4 Organization of the Book.- 2. Survey Of Contemporary Content-Based Image Retrieval Systems.- 2.1 Systems Using Edge Points.- 2.2 Systems Using Color Histograms.- 2.3 Systems Using Textures.- 2.4 Summary.- 2.5 Systems Using Object Regions.- 2.6 Systems Using Other Features.- 2.7 Comparison of the Image Features.- 3. Indexing Structures For Image Databases.- 3.1 KDB-Tree.- 3.2 R-Tree.- 3.3 R*-Tree.- 3.4 SS-Tree.- 3.5 SR-Tree.- 3.6 Comparisons.- 4. Building An Advanced Region-Based Image Retrieval System (ARBIRS).- 4.1 The System Goal.- 4.2 Feature Space Selection.- 4.3 The Challenges.- 4.4 The System Outline.- 5. The Texture Detection Subsystem.- 5.1 Separating Textures From Non-Texture Regions.- 5.2 Examples of The Texture Detection.- 6. The Region-Based Subsystem.- 6.1 Segmenting Images Under Non-Uniform Illumination.- 6.2 Segmentation Using Human Perceptual Dimensions.- 6.3 The Mathematical Model.- 6.4 Outline of the Segmentation Method.- 6.5 Segmenting Chromatic Colors.- 6.6 Segmenting Achromatic Colors.- 6.7 Detection of the Linear Chroma-Value Correlation.- 6.8 Experimental Results.- 6.9 The Indexing Scheme.- 7. The Histogram-Based Subsystem.- 7.1 Problems of Traditional Histograms.- 7.2 Color Histogram Creation.- 7.3 The Indexing Scheme Based on the Color Histogram.- 8. The Database Query Subsystem.- 8.1 Database Query by Simple Regions.- 8.2 Database Query by Texture Regions.- 8.3 Database Query by Compound Regions.- 8.4 The User Interface.- 9. Experimental Evaluations and Discussions.- 9.1 Evaluations.- 9.2 Discussions.- References.