Key issues in video summarization and its application

The summarization of video content provides an effective way to speed up video browser and assist for video retrieval, this paper proposes a novel method for automatically summarizing multi-view video contents. Three kinds of content views such as semantic-face, video caption and visual features are presented in this paper. First, the human faces in video and images bear significant semantics, thus we can summarize video content by indexing and analyzing video faces, especially for video news. Second, video caption could be used to generate video summarization with high-level semantics since it implied lots of semantics inherently. Finally, the similar video shots could be integrated into concise representation through clustering and mining of visual features.