一种新的基于RS理论的压缩域镜头分割算法  

Novel compressed-domain shot segment alogrithm based on RS theory

作　　者：李向伟[1,2] 李战明[1] 张明新[3] 张国权[1] 韦哲[1] 

机构地区：[1]兰州理工大学电气工程与信息工程学院,兰州730050 [2]兰州工业高等专科学校,兰州730050 [3]西北师范大学数学与信息科学学院,兰州730070

出　　处：《计算机应用研究》2009年第4期1588-1590,共3页Application Research of Computers

基　　金：甘肃省自然科学基金资助项目(3ZS051-A25-047)

摘　　要：针对传统的非压缩域镜头分割算法数据多、运算量大和效率低的缺点,提出了一种基于RS理论的压缩域镜头分割算法。该算法首先根据MPEG压缩标准,从视频流中提取DCT系数;然后经预处理得到每一帧的DC系数;最后依DC系数建立镜头分割信息系统模型,通过RS理论的划分与属性约简得到视频镜头。实验结果表明,该算法相对传统非压缩域算法运算量和数据量大大减少,算法效率明显提高。Aimed at overcoming the amount of computing and low efficient of traditional shot segment methods in noncom- pressed-domain. The paper presented a new compressed-domain shot segment algorithm based on RS theory. The method ex- tracted DCT coefficients from the video stream according to MPEG standard firstly. And then got the DC coefficients after the preproeessing. At last, constructed the information system using DC coefficients and got shot segment model. Experimental re- sults show that the alogrithm reduced the amount of computing and data obviously, the efficiency is increased dramaticly.

关 键 词：镜头 镜头分割 RS I帧 压缩域 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]