A BSU-based sports video content analysis framework for content-based streams filtering

The flowing of a good deal of sports video, the widespread application of Internet and the rapid increase of the users' requirements on demand make content-based sports video streams filtering important and urgent. Sports video content analysis module is the base of the content-based video streams filtering system. In this paper, we propose a novel and general sports video content analysis framework for content-based sports video streams filtering. The framework is based on Basic Semantic Unit (BSU) that is specially presented for sports video in this paper. And we demonstrate the application of this framework on Soccer Video. Our preliminary experimental studies indicate that the proposed framework is satisfactory and promising.