一种基于草图注释的视频浏览技术

Sketch Annotation Based Video Browsing

针对视频的高效浏览和定位问题,提出一种基于草图注释的视频浏览技术.首先根据视频流中各帧的特征点提取关键帧;然后利用草图为视频的主要内容添加静态和动态注释,并结合提取的关键帧生成表征视频内容的个性化视频信息库;最后定义草图手势库来进行相关的视频交互操作.实验结果表明,该技术可有效地组织视频语义信息,提高视频的浏览、理解与交互的效率,并具有较高的用户满意度.

To browse and locate video efficiently, a sketch-based video annotation and browsing technique is proposed in this paper. First, key frames are extracted according to the features in the video, and static and dynamic sketch annotations are interactively specified for the video. Then a personalized sketch-library is created based on the key frames and interactive sketch annotations. Finally, a sketch based interface via gesture commands is proposed to perform interactive video operations. Experimental results show that the proposed sketch annotation and interaction can improve the efficiency of video browsing, understanding and interaction, and brings better user experiences.