Multimedia and the Semantic Web

The growing interest in digitising and making available media items has resulted in an ever increasing amount of diverse media available on the Internet. These media objects come from many domains, in many formats, have varying levels of available, structured metadata and form a valuable, under-utilised resource.

The semantic web seeks to make content machine understandable and so uses metadata and defined semantic knowledge structures to support access and retrieval of information and to enhance interoperability between different sources of knowledge.

There is a growing desire across many communities to 'do something' with media repositories beyond making them available via a simple web search interface. Museums, libraries, educational and scientific communities are interested in services such as

• annotation (both automatic and tool supported)
• inference of relationships between atomic media items across systems and domains
• automatic generation of personalised multimedia presentations

The aim of this project is to develop ontologies, standards, tools and techniques to enable the incorporation and exploitation of multimedia resources with the semantic web. This will potentially involve construction of efficient knowledge-based multimedia systems that automatically extract, from a multimedia input, semantic information (objects, events, properties, relations) described in ontologies of specific domains enriched with audiovisual data. The extracted semantic metadata can be used for classification, summarisation, indexing, searching and efficient retrieval of multimedia content.

This project is being undertaken by University of Queensland, ITEE Dept. PhD student, Suzanne Little under the supervision of Dr Jane Hunter within the MAENAD group at DSTC.

The name sunago (pronounced soon-ag'-o) is ancient Greek loosely meaning to gather together, combine or collect and infer or build new meaning.