Multimedia

Abstract

This summary overviews a keynote talk that the author is giving at the WebMedia conference. Christel will discuss his journey with multimedia research over the past six years, taking him from Carnegie Mellon University's Computer Science Department into the Entertainment Technology Center (ETC). The story begins with the use of speech recognition, image processing, and language technologies to automatically process large video corpora. Such processing facilitates more efficient retrieval. As demonstrated by top scores in the international TRECVID benchmarking forum, Christel's CMU Informedia research group has experienced success in finding relevant video shots quickly from large masses of material. The emphasis is on leveraging the intelligence of a human user in the interactive retrieval loop, with lessons shared in visual analytics papers.