Efficient database support for WWW image retrieval

Abstract

WWW is exploding and shaping the current research direction. To enhance the WWW apge content, images are increasingly being embedded in HTML docuemnts. Such documents over the WWW essentially provide a rich and interesting source of image collection from which users can query. In this thesis, we first propose an effective semantic-based WWW image retrieval system, and extend it with multiple visual features. To provide efficient database support, we then study the problem of high-dimensional indexing, from which we further address the problems of hyper-dimensional indexing and multiple high-dimensional indexing.