摘要
视频是一种非格式化、时空并存、大数据量的媒体数据,传统的基于文本的信息查询技术已不能满足人们的要求,所以有关视频特征描述、查询、浏览及管理中存在的障碍亟需解决.本文在基于内容的视频检索系统中提出了一种消隐镜头的检测算法———基于局部特征时变一致性的检测算法.同时利用MPEG-7标准中描述符的结构定义描述视频检索系统的特征库,实验结果验证了系统的有效性.
Video is non-formatted, space-time combined and with a great deal of data. For such media, the traditional text-based information query system can not meet our needs. In order to efficiently make use of video information, some issues, such as feature representation, querying, browsing and management badly need to be solved. This paper proposes an algorithm called local coherent temporal change to detect dissolved shot, and the descriptors in MPEG-7 are used to describe feature database in our CBVR system. Experimental results show the efficiency of the system.
出处
《北京交通大学学报》
EI
CAS
CSCD
北大核心
2005年第5期79-83,共5页
JOURNAL OF BEIJING JIAOTONG UNIVERSITY
基金
国家自然科学基金资助项目(60373028)
教育部博士点专项基金
新世纪优秀人才支持计划资助项目