Characterization of image content

Abstract

With the recent developments in Internet Web-based applications, the requirements in international access has evolved from one which is command-based for direct manipulation of data, and that which is text-based to one which is visual-based. Most significantly, multimedia applications which require the support of visual information access have grown exponentially in number. Coupled with the evolving paradigm in visual-oriented access and the growing prevalence of digital images and videos the need for more effective means of image retrieval based on visual content is increasing more important. Traditionally, access to visual information using text-based indexes, which are built upon human supplied annotations is ineffective. Visual information in image and video libraries requires new algorithms for automatic extraction and indexing of relevant image features.