| 基于场景的视频摘要生成方法 |
| |
| 引用本文： | 智敏.基于场景的视频摘要生成方法[J].内蒙古大学学报(自然科学版),2009,40(1). |
| |
| 作者姓名： | 智敏 |
| |
| 作者单位： | 内蒙古师范大学计算机与信息工程学院,呼和浩特,010022 |
| |
| 摘 要： | 针对如何在语义层次上形成视频摘要问题,提出了一种基于场景的视频摘要生成算法.首先利用类内和类间损失对场景进行分类,然后根据语义概念"重要度"从镜头类和场景提取视频摘要的构成帧,最后将选取的关键帧根据时间顺序排序,构成静态浏览型的视频摘要.实验证明该方法可以用少量的视频帧有效地表示视频的主要内容,并能有效地保持视频内容的连续性.
|
| 关 键 词： | 视频摘要 语义 重要度 |
A Scene-based Video Summary Generation Algorithm |
| |
| ZHI Min.A Scene-based Video Summary Generation Algorithm[J].Acta Scientiarum Naturalium Universitatis Neimongol,2009,40(1). |
| |
| Authors: | ZHI Min |
| |
| Institution: | College of Computer and Information Engineering;Inner Mongolia Normal University;Hohhot 010022;China |
| |
| Abstract: | A scene-based video summary generation algorithm in semantic level is presented.The scenes are classified by intra-class and inter-class losses and the frames constructing the video summary are selected from the key frames of every scene by the introduced importance function.Experimental results show that the summary generated by the algorithm can properly represent the main contents of a video,and effectively keep the continuity of video content. |
| |
| Keywords: | video summary semanteme importance measure |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
