Achievements and remaining challenges in multimedia data modeling

Summary form only given. Multimedia research over the past decade has resulted in the development of multimedia data managers that have achieved a reasonable level of sophistication. It is possible now to organize images and video data efficiently, model them reasonably well and perform searches on them based on selected features (query-by-example). The development of XML has facilitated the storage and retrieval of structured text. We have, however, not yet fully addressed the more difficult problems: how to capture and model the semantics of the media objects, how to query using these semantic properties, how to exploit the interrelationships among media (how to develop fully multimedia systems), how to architect these systems, and others. In this talk, we review where we have come and then discuss some of the remaining challenges