计算机工程 ›› 2010, Vol. 36 ›› Issue (7): 227-229.doi: 10.3969/j.issn.1000-3428.2010.07.078

• 多媒体技术及应用 • 上一篇    下一篇

基于内容的视频拷贝检测研究

刘 红1,2,文朝晖2,王 晔1,2   

  1. (1. 第二军医大学网络信息中心,上海 2. 复旦大学计算机科学技术学院,上海 200433)
  • 收稿日期:1900-01-01 修回日期:1900-01-01 出版日期:2010-04-05 发布日期:2010-04-05

Study on Content-based Video Copy Detection

LIU Hong1,2, WEN Zhao-hui2, WANG Ye1,2   

  1. (1. Network Information Center, Second Military Medical University, Shanghai 200433; 2. School of Computer Science, Fudan University, Shanghai 200433)
  • Received:1900-01-01 Revised:1900-01-01 Online:2010-04-05 Published:2010-04-05

摘要: 提出基于图的视频拷贝检测方法,该方法将视频序列匹配结果转换为匹配结果图,进而将视频拷贝检测转换成在匹配结果图中查找最长路径的问题。实验结果显示基于图的序列匹配算法拷贝定位准确度高,可弥补图像底层特征描述力不足的缺陷,节约检测时间,批量定位2段视频序列中可能存在的多段拷贝。

关键词: 视频拷贝检测, 聚类, 最长路径

Abstract: This paper proposes an efficient graph-based approach for video copy detection. It converts the video sequence matching results to a matching results graph, so the problem of video copy detection becomes a problem of finding the longest path in the matching results graph. Experimental results show that the graph-based video sequence matching algorithm has many advantages, such as high copy locating accuracy, making up for inadequate description of image low-level features, saving detecting time, batch detecting the more than one copy which exist in two video sequences and so on.

Key words: video copy detection, clustering, longest path

中图分类号: