The THISL spoken document retrieval project

Abstract

THISL is an ESPRIT Long Term Research Project focused on the automatic indexing and retrieval of broadcast television and radio programmes. In particular it is concerned with the production of a demonstrator news-on-demand system to navigate an archive of BBC news broadcasts. Prototype systems based on both British and North American broadcast news have been constructed. The North American system has been successfully evaluated within the framework of the TREC-6 and TREC-7 spoken document retrieval tracks, and the system based on BBC TV and radio news archives will be evaluated by BBC R&D.