一种基于内容的镜头边界快速检测方法

A Method for Fast Shot Boundary Detection Based on Content

镜头边界检测是视频索引研究领域中的一项重要技术和基础步骤,近年来受到广泛的研究,但大量的研究过度关注算法的检测精度(从查全、查准两方面考虑)而忽略了算法的运算成本.就目前已有的框架和算法进行改进,提出一种兼顾检测精度和运算速度的镜头边界检测方法.该方法引入预处理技术,在进行镜头边界检测前过滤掉视频中不包含镜头转换的视频序列,从而在不失精度的前提下提高算法运行速度.试验证明该方法在提高检测速度的同时可以有效地检测出视频序列中的突变、渐变边界,并且获得了令人满意的精度.

As the important technic and fundamental step of video retrieval,shot boundary detection has been widely researched in recent years.However,most of the researches just focus on achieving excellent detection accuracy while the high computational cost has been disregarded.A fast shot boundary detection algorithm is presented,which improved from exiting frameworks and algorithms.With the use of pre-processing,the speed of detection can be largely increased without losing accuracies,because a large number of non-boundary video segments have been discarded before boundary detecting.Experimental results indicate that the developed algorithm can achieve excellent detection accuracy while effectively accelerate the shot detection process.