基于内容的视频检索与挖掘关键技术研究

The research of video retrieval techniques based on content

视频检索与挖掘是当前计算机视觉、视频分析及知识挖掘等领域的研究热点之一,在现实生活中有广泛的应用前景。基于内容的视频检索与挖掘是一种自动、智能的多媒体信息检索技术,借助于最新数学分析理论与工具,对视频数据进行建模,在此基础上进行智能分析处理的一项技术。本文从分析基于内容的视频检索与挖掘系统的结构和原理出发,详细的讨论了视频检索与挖掘的关键技术:镜头检测与关键帧提取,并对新的国际标准MPEG及其框架下的视频检索与挖掘技术作了介绍,最后对视频检索与挖掘的前景与发展方向作了展望。

Video retrieval and mining is the research hot in computer vision, video analysis and knowledge mining in current. It has a wide application prospect to reality. The content based video retrieval and mining is a multimedia information retrieval techniques using automated, intelligent methods, and by means of novel mathematical theory and tools, constructing the data model to the video information, and based which many intelligent analysis techniques are performed. The paper stated with the basic structure and principle of video retrieval and mining based on content, specifically discussed the key techniques:shot detection and key frame extraction. At the end of paper, the new international standard MPEG and the video retrieval and mining techniques in MPEG mainframe is introduced, meanwhile, the development and prospect of which are also presented.