A Hierarchical Description Method for Video Content

摘要

This paper proposes a new method of video content description, which will be used in video content-based retrieval. The method proposed in this paper has a hierarchical video description structure. The hierarchical description structure includes the aspects from low levels to high levels. It is feasible to describe the visual features of objects in a video by this means. Three aspects of video content description in this paper are investigated, such as video description, the visual features of video content extraction and database search engine. Motion information is used to segment objects. Subspace searching is used in the search engine. The video content-based retrieval experiment system is introduced to examine the effect of the research work.