| 摘要 |
PROBLEM TO BE SOLVED: To provide a video analysis technology for more efficiently managing and handling video content. SOLUTION: For example, highlight extraction is fundamentally carried out in three layers of a layer 1 to a layer 3. The layer 1 is an intershot shot type temporal relevance rule, that is, a layer that represents a statistical model showing such a relevance rule. The layer 2 is a layer that represents an interframe temporal relevance rule having meaning among frames constituting a prescribed shot, The layer 3 is a layer that clarifies meaning of the frame by feature quantity in the frame. That is, a layer that represents a relevance rule of each frame in the shot, whose meaning is clarified by the layer 3 is the layer 2. This invention is applicable to a field that handles video content. COPYRIGHT: (C)2008,JPO&INPIT
|
