Vidipedia: Community Indexing for Audiovisual Material

Video is the most popular media format, and the potential value hidden in the tens of millions of hours of footage in the world’s media archives is enormous. However, such a great volume of footage also means that it is very difficult to search within it for anything specific. In order to facilitate access, archive content needs to be widely spread, which requires digitisation. Because the process of digitisation is costly, archive owners are unwilling to commit to it until a) a cost-effective mechanism exists to annotate the collection in such a way that potential users can easily search audio and video content, and b) there is a working business model that justifies the cost of digitisation by demonstrating new revenue streams that will appear as a result of making the collection available.

The aim of the Vidipedia project was to address these needs by examining the potential for community based annotation and identifying a business model that supported it. The project intended to create a tool that would address the challenges of archiving, search and discovery for producers and consumers of multimedia content. Another intention of the Vidipedia project was to enable interoperability at the semantic level between services and systems that would support inter-enterprise applications.

Date: Oct 2006 – Jan 2009

Funding: DTI  £81,515

Researchers: Patrick Olivier (PI), Paul Watson (CI) – School of Computing Science, Dan Jackson.

Collaborators:  ITV Tyne Tees, University of Sunderland, Enigma Interactive