基于视频摘要生成技术的研究  

Research on Video Abstraction Techniques

作　　者：朱志辉[1] 

机构地区：[1]湖南商务职业技术学院

出　　处：《微电子学与计算机》2006年第2期76-78,82,共4页Microelectronics & Computer

摘　　要：文章研究了标题形式摘要,故事板摘要及缩略视频摘要三种形式的摘要。充分利用各种多媒体融合分析手段,提出了视频内容判定模型。根据不同的视频分解粒度,提出了不同层次的对象重要度判定模型,生成有意义的视频摘要。设计实现了有效的视频摘要生成系统,融合多种技术与方法,形成完整的检索视频索引生成系统。During recent years, more and more multimedia applications have changed from style of playing to style of content-based access or retrieval and interaction. Rapid transaction of video-media by Internet is one of the basic problems. The content of video is most rich and it＇s style is most multiplex, more and more people like this media. But videos are linear and the data of them is huge. Now the speed of media is increasing rapidly, but band of Internet is limited. So how to find a method to retrieve some interested segments is the most important thing. This paper probes on the video construction and contend based analysis which assist video analysis. Work over technology of video Summarization. Study other algorithm and analysis the feature and technology of them, we propose our own method, which make best use of multimedia fusing analysis method. We propose a judgment model to testify video content. We create video Summarization by our method fully use of video construct information, artificial text and length of shot.A video Summarization creating demo-system is proposed.

关 键 词：新闻视频 视频摘要 字幕探测 

分 类 号：TP37[自动化与计算机技术—计算机系统结构]