Using Image Content for retrieval and Navigation

Multimedia information collections are becoming increasingly common in the humanities. This paper describes developments in hypermedia and image analysis techniques which are beginning to provide tools for retrieving and navigating through digitally stored visual information. The use of static links for navigating from text has become increasingly popular with the emergence of the World Wide Web, but generic links, based on information content matching can provide more powerful information handling capabilities. We show how content based retrieval and generic link authoring and following can be applied to image data in multimedia collections. The processes at present rely on matching image selections using basic image features such as colour, texture and outline shapes. The paper highlights some of the problems with content based techniques applied to image retrieval and navigation and suggests how the introduction of a multimedia thesaurus could be introduced to overcome some of these problems. The multimedia thesaurus provides a network of concepts in the domain of interest