摘要
如何在海量的图像、视频和音频数据中快速找到用户关心的内容是检索领域研究的热点之一。从体育视频结构的特点出发,分析并给出一种语义标注及分层索引方法,对于一个体育视频数据,可通过与特征库匹配自动完成视频数据标注。检索时根据分层分级结构索引,能快速定位搜索范围,并根据用户反馈信息,执行一个新的动态学习检索过程,提高检索效率。
How to rapidly find the contents which users caring about in massive data of images, video and audio is one of the hot topics in information retrieval area. In this paper, proceeding from the features of sport video structure, we analyse and give a semantic tagging and hierarchical indexing method. For a sport video data, based on matching with the feature database, the video data tagging can be automatically completed. When retrieving, the search range can be fast fixed using hierarchical classification structure index. Moreover, combining with user＇ s feedback, the method starts a new dynamic retrieving procedure to improve the retrieving performance.
出处
《计算机应用与软件》
CSCD
北大核心
2012年第10期258-260,共3页
Computer Applications and Software
基金
西藏民族学院青年学人培育计划资助项目成果(12myQP07)
关键词
体育视频
分层语义
台标
分层索引
Sport video Hierarchical semantics TV symbol Hierarchical index
