The University of British Columbia

Summarization of video content is necessary in order to reduce the large amount of data involved in video retrieval. In a frame-based digital video retrieval framework, this is achieved by representing the content of a video sequence using key frames. Similarly, in an object-based framework, such as the one suggested by the MPEG-4 standard, video object planes can be used for summarization of video object content. We propose a method for key video object plane selection using the shape information in the MPEG-4 compressed domain. Two popular shape distance measures, the Hamming and Hausdorff distance measures, are employed to measure the similarities between the approximated shapes of the video objects. The corresponding algorithms, with different implementation complexity and computation tradeoffs, select key video object planes that represent efficiently the salient content of the video objects