摘要
基于内容的多媒体检索特别是视频检索,由于多媒体数据本身具有复杂的语义,所以极大地提高了检索的难度。算法着眼于视频本身挖掘出充分的资源信息并且将这些信息加以融合来提高视频检索的性能。基于这种思想,提出一种多模态视频检索模型以及相应的手动式搜索和交互式搜索的算法方案。搜索策略在TRECVID视频检索比赛中取得了不错的成绩。
Complicated semantics contained in multimedia data significantly increases the difficulty of content based multimedia retrieval, especially video retrieval, The algorithm we presented focuses on mining ample resource information from video itself and fusing the,~ to im- prove the performance of video retrieval. We propose a muhimodal retrieval model in this paper and introduce the corresponding algorithms of manual search and interactive search. Our searching strategy performed well in TRECVID video retrieval contest.
出处
《计算机应用与软件》
CSCD
2009年第8期205-208,共4页
Computer Applications and Software
