Media semantics: who needs it and why?

Introduction As pointed out in the keynote address at the 2001 ACM Multimedia Conference [3], the current major goal of multimedia research is directed towards provisioning information for pervasive access and use. To achieve this, what will become important are technologies that help sift useful nuggets of information from torrents of media data, which can be turned into valuable knowledge just in time and need, and tools that help provide access to these nuggets in anytime anywhere any device mode to everyone ranging from enterprise customers to independent consumers. Further we need to treat various media on an equal basis in environments that provide multimedia-based interactions, where they ultimately add value to users, whatever the nature of the interactions may be and whatever the preferred mode of media access may be. Thus, there is a fundamental need to investigate the means to elucidate, sublimate, or rationalize information and knowledge from media data. However, current user expectations are far from being met owing to generic low-level content metadata available from automated processing that deal only with representing perceived content, and not the semantics of it.