A Hierarchical Video Annotation System

Semantic search service is generally considered to include text search, audio search, and video search. The annotation of a video is crucial to the success of video search and retrieval. With the precise annotation of video files and their semantic reasoning, more accurate search results according to the users´ query intention can be obtained. In this paper, a prototype of such a video annotation system, called Semantic Video Annotation System (SVAS), is presented. SVAS uses a three-level annotation architecture and a semantic video search language called Semantic Query Description Language for Video (SQDL-V). SQDL-V engine based on SVAS is able to return more accurate search results in comparison to the formal video search method.