We proposes "Video Icon Diagram(VID)"for representing inner structures of a video. VID is a graphical diagram composed of video icons each of which illustrates a video segment, such as a shot. The icons and their layout represent semantics relations such as order, hierarchy, or equivalence, and combination of those structures represent the semantic structure of a video. In this paper, we first propose our framework and the methods for generating a VID by making correspondence between a scenario tagged by XML and a video. Our experiments shows the effectiveness of our method.
展开▼