摘要
提出了一个基于口播检测的高效新闻视频单元切分方法。该方法首先检测出新闻视频的镜头边界;然后从每个镜头中提取出关键帧,并计算出关键帧的直方图和SIFT特征;最后通过关键帧聚类获取新闻视频中的所有口播镜头,并以此为依据将新闻视频分割成多个语义单元。基于以上方法,开发了用于新闻视频单元切分的软件系统。该系统能够准确、高效地实现新闻单元的自动切分,有效地减轻视频切分时的工作强度,满足新媒体时代节目快速制作的要求。
An efficient method is proposed to segment news videos based on anchorperson detection. Firstly, the shot boundary of a news video is detected. Then, the keyframe of each shot is extracted, and their histogram and SIFT features are calculated. Finally, all anchorperson shots by clustering the keyframes are retrieved, and the news video is segmented into several semantic fragments. Based on the method above, a software system is developed, which can automatically segment news videos accurately and efficiently. This system can greatly reduce users＇ labor in news video segmentation, and meet the requirement of fast production in new media times.
出处
《电视技术》
北大核心
2014年第23期170-173,共4页
Video Engineering
基金
浙江省自然科学基金项目(LY14F020050)
关键词
新闻单元切分
口播检测
直方图
聚类
segment news videos
anchorperson detection
histogram
clustering
