改进相关反馈技术在CBVR人体动作识别中的应用研究

The application and research of improved relevance feedback technology in CBVR human action recognition

针对CBVR人体动作识别问题,提出利用相关反馈技术与其它技术相结合。首先,分析了从视频中提取人类行为并将其表示成特征集的方法;然后研究了各种基于内容的视频检索方法;最后,将这些技术进行组合得到基于相关反馈技术的最优CBVR人体动作识别方法。在三个行为数据库中(包括UCF运动,UCF视频网站和HOHA2)评估了由上述几种技术组合而成的性能,实验结果为开发高效的基于内容的视频检索系统提供了有益的借鉴。

For the issue of CBVR human action recognition,relevance feedback technology is proposed to be applied to fusion with several technologies. Firstly,various ways that an action can be extracted from a video and represented as a set of features are analyzed. Then various content- based retrieval methods are presented. Finally,the technologies above are combined to form the best CBVR human action recognition method based on relevance feedback technology. The performance of several combinations of the above techniques in three realistic action datasets： UCF Sports,UCF You Tube and HOHA2 are evaluated; and the experiments can provide references for exploring content-based video retrieval system.