The invention discloses a method and a system for cataloging news video. The method is based on the caption strip, the host and the audio mute point information in the news program to automatically catalog the news video. News video streaming audio and video to audio data separation, music, news, to determine the effective time range in a file; within the valid time range to determine the time point, the host audio mute frame and caption frame, and carries on the comprehensive analysis, to determine the news item segmentation time points; recognition of video caption information in association with the segmentation results as the semantic information cataloging. The system includes: the strip module and the export module, which are connected with the news video strip result database. The browsing module, the playback module and the correction module are connected in parallel between the user and the news video strip result database. The invention solves the problem that news automatic strip cutting and automatic semantic information annotation of news items are realized, and automatic cataloging of news programs is realized, and the utility model has the advantages of high efficiency and low cost.
【技术实现步骤摘要】
【技术保护点】
一种新闻视频编目方法,其特征在于,基于新闻节目中出现的字幕条、主持人、音频静音点信息对新闻视频进行自动编目,步骤如下: 步骤1:对新闻视频流进行音视频数据分离,获得音频数据和视频数据; 步骤2:对音频数据进行片头音乐匹配,确定新 闻节目在文件中的有效时间范围;对新闻节目所在时间范围内音频数据进行静音点检测,获得音频静音点序列;对新闻节目所在时间范围内视频数据进行关键帧抽取、主持人图像帧检测和文字帧检测,获得新闻节目所在时间范围内的静音点时间、主持人出现时间、文字出现时间; 步骤3:对音频静音点序列、主持人出现时间和文字信息出现时间进行综合分析处理,获得 ...
【技术特征摘要】
【专利技术属性】
技术研发人员:陈众,张树武,曾智,杨武夷,
申请(专利权)人:中国科学院自动化研究所,
类型:发明
国别省市:11[中国|北京]
还没有人留言评论。发表了对其他浏览者有用的留言会获得科技券。