一种新的基于I帧的压缩域镜头分割算法  

A New Compressed Domain Shot Segment Method Based on I Frame

作　　者：张国权[1] 李向伟[1] 李战明[1] 张明新[2] 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 [2]西北师范大学数学与信息科学学院,甘肃兰州730070

出　　处：《海南大学学报（自然科学版）》2008年第3期273-277,共5页Natural Science Journal of Hainan University

基　　金：甘肃省自然科学基金(3ZS051-A25-047)

摘　　要：针对传统的非压缩域镜头分割算法数据量大、运算量大和效率低的问题,提出了一种基于I帧的压缩域镜头分割算法.该算法首先根据MPEG-2压缩标准,从视频流中提取DCT系数,然后经过预处理得到每一帧的DC系数,最后依DC系数建立信息系统并得到镜头分割模型.实验结果表明,此算法相对于传统的非压缩域算法,可大大减少运算量和数据量,提高算法的效率.Aimed at overcoming the amount of computing and low efficiency of traditional shot segment methods in non-compressed domain. This paper presents a new compressed domain shot segment method based on Ⅰ frame. The method extracts coefficients of DCT from the video stream according to MPEG standard firstly. And then it gets the DC coefficients after the preprocessing. At last, it constructs the information system using DC coefficients and gets shot segment model. Experimental results show that the method can reduce the amount of computing and data obviously and it results in increasing efficiency, compared with traditional shot segment methods.

关 键 词：镜头 镜头分割 Ⅰ帧 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]