• 中国计算机学会会刊
  • 中国科技核心期刊
  • 中文核心期刊

J4 ›› 2008, Vol. 30 ›› Issue (3): 40-43.

• 论文 • 上一篇    下一篇

基于综合相似度的二次差分法在镜头检测中的应用

叶军[1] 吴菲[2] 李建良[3]   

  • 出版日期:2008-03-01 发布日期:2010-05-19

  • Online:2008-03-01 Published:2010-05-19

摘要:

本文提出了一种基于综合相似度的二次差分法,综合考虑小波分析和统计量等多种方法对镜头进行检测。在镜头突变检测中考虑了相邻帧的二次差分法;同时,在渐变检测中引入了自适应阈值法,考虑了非相邻帧的二次差分法,避免了人为因素,克服了其他方法仅适用于一类视频序列或几类视频序列的限制,因而具有较好的鲁棒性。实际视频数数据的实验结果表明,本文所用方法对镜头边界检测具有很好的效果。

关键词: 镜头边界检测 帧分块 小波变换 似然比 综合相似度

Abstract:

An important step in content-based video retrieval is the temporal segmentation of video. There are some basic algorithms , including the histogram algorithm,the double-thresholding method,which are used to detect abrupt shot changes and gradual shot changes, but they cannot always attain high-precision detection. We propose a twi-difference algorithm for shot boundary detection based on the similarity that integrates the wavelet analysis and statist ics of frames. And the algorithm is used to detect the abrupt shot changes and gradual shot changes based on the adjacent frames and the non-adjacent fr ames. It is proved that the twi-difference algorithm obviously improves the detecting precision.

Key words: shot boundary detection, frame blocking, wavelet change, ratio rate. integrated similarity