New Frontiers for Intelligent Content-Based Retrieval

Samenvatting

New developments in intelligent infrastructure-based content retrieval systems are explored in this article. In this context, “intelligence” is defined as a system’s capacity to reason and learn effectively, use context to its advantage, construct and maintain models of the world or specific situations, and make use of dynamic representations of information. In order to build efficient systems for retrieving audiovisual material at semantic levels comparable to human perception and cognition, we contend that these components are necessary. We survey studies from cognitive psychology, AI, semiotics, and computer vision that provide light on the nature of intelligence and how intelligent systems are built. We go on to talk about how content-based retrieval systems might take use of new possibilities made possible by some of the key concepts in these areas. Lastly, we highlight a few 
of our initiatives in these areas. A multimedia knowledge presentation system called Media Net and some MPEG-7 description tools that allow intelligent content-based retrieval are specifically introduced