Implementing a Content-Based Public-Oriented Audio and Video News Retrieval System

摘要

Video is poised to largely replace both text and images as the media for transmitting information inrnthe coming years. The challenge of the Information Processing community is how to index therninformation found in this voluminous and dynamic media stream. Most of the linguisticrninformation is encoded in the audio channel of video data, which, once transcribed, can be accessedrnusing text-based tools. This talk will describe our current research in providing an index into therncontent of video and audio streams, using LIMSI’s state-of-the-art automatic speech transcriptionrnsystem for French, English, Mandarin and Arabic languages. I will also describe and demonstraternthe Voxalead News system, and other results of the French-German Quaero project, that integraternresults from industry and research, for the next generation of video-based information searching.