Semantic Analysis and Retrieval of Sports Video

For bridging the semantic gap, specific-domain approaches are employed in our system, a fully automatic and computationally efficient framework for the indexing, browsing and retrieval of sports video is proposed. A hierarchical organizing model based-on MPEG-7 is introduced to effectively represent high-level and low-level information in sports video. The model's attributes have been instantiated through a series of multi-modal processing approaches for soccer video. To effectively retrieve highlight clip of sport video, adaptive abstraction is designed based on the excitement time curve, and an XML-based query scheme for semantic search and retrieval is proposed based on the hierarchical model. We develop a robust detection algorithm to automatically locate the field region of sports video based on low-level color features. An effective mid-level shot classification algorithm is proposed based on spatial features and the coarseness of human figures. For retrieval, XQuery, a XML query language, is employed in the proposed querying framework. The proposed framework employs visual, audio, and cinematic features, and can be scalable to different preferences and requirements. Its efficiency, effectiveness, and the robustness of the proposed framework are demonstrated over our extensive experiments