Computer Engineering and Applications ›› 2009, Vol. 45 ›› Issue (31): 233-236.DOI: 10.3778/j.issn.1002-8331.2009.31.070

• 工程与应用 • Previous Articles     Next Articles

New method for news events segmentation based on multiple features

LIU Wen-ping1,2,JIANG Xiao-sen2   

  1. 1.Department of Digital Media,Beijing Forestry University,Beijing 100081,China
    2.Department of Computer Science,North China University of Technology,Beijing 100041,China
  • Received:2008-06-12 Revised:2008-10-19 Online:2009-11-01 Published:2009-11-01
  • Contact: LIU Wen-ping



  1. 1.北京林业大学 信息学院 数字媒体教研室,北京100083
    2.北方工业大学 计算机系,北京 100041
  • 通讯作者: 刘文萍

Abstract: A new method based on video,audio and text information for news events segmentation is proposed.A prototype system for the news events segmentation,browsing and retrieval is then developped for testing the proposed method.In this new approach,all the caption frames,anchorperson frames,silence gap and speaker change in the news video are detected and used for segmenting the news events.The experimental results demonstrate that the proposed method has high segmentation precision than that of the caption-based method.

Key words: video retrieval, news events segmentation, caption extraction, speaker change detection, silence gap detection

摘要: 提出了一种基于新闻视频中的标题字幕信息和音视频特征对新闻事件进行分割的方法,并实现了一个新闻事件分割、浏览和检索的原型系统。提出的方法综合利用新闻视频中的标题检测、主持人画面检测以及静音片段和语者切换检测技术分割整段新闻中的新闻事件。实验结果表明,该方法较仅利用标题的新闻事件分割方法在分割准确性上有了显著提高。

关键词: 视频检索, 新闻事件分割, 文字提取, 语者切换检测, 静音片段检测

CLC Number: