【摘要】 <正> 1.引言在当今的信息社会里,多媒体技术、计算机网络技术的迅速发展使得多媒体技术的应用也随之迅速发展。然而,可视化信息的大量生产和传播也带来一个问题.即面对信息量如此浩大的多媒体数据库,用户应如何找到自己需要的信息。例如:用户可能提出这样的要求,“找出数据库中所有赛车的场面”。我们需要研究的是,信息检索系统应如何适当表达这个要求的内容,并在数据库所有的视频信息中找到符合这个查询要求的信息返回给用户。于是,大规模多媒体数据库的索引和检索具有的巨大实用价值使得它在国外已成为一个重要的科研课题。更多还原

【Abstract】 While large video/multimedia databases became increasing common, traditional retrieval methods are no longer adequate for video data. Since video information is temporal,spatial,unstructured and massive,the new content-based video retrieval,which can retrieve databases by the video content,will supply user better. We can classify Video CBR into two categories:the improved traditional methods and the visual methods. In this paper we describe the main idea and recent studies on video CBR. After analyzing different development models of some typical video CBR system-s,we point out their distinguished implement characters and shortcomings. At last,we suggest some issues that will improve query efficiency based on forecasting the trend of video retrieval technique.更多还原