Content-based image indexing and retrieval using computational intelligence

Abstract

The significant growth in the volume of image data has driven the demand for efficient techniques to index and access the image collections. These techniques are used in fields including applications such as online image libraries, e-commerce, biomedicine, military and education, among others. In view of this, content-based image retrieval (CBIR) has beendeveloped as a scheme for managing, searching, filtering, and retrieving the image collections. CBIR is a process of retrieving a set of desired images from the database on the basis of visual content such as color, texture, shape, and spatial relationship that are present in the images. The problem is challenging due to the semantic gap between the low-level visual features and the high-level human perception. With the objective to reduce the semantic gap, this thesis investigates several challenging problems in current CBIR systems. It covers the following three main aspects: relevance feedback in CBIR (Chapters 4 and 5), relevance feedback in region-based image retrieval (Chapter 6), and peer tagging and knowledge propagation (Chapter 7).