一种基于背景提取的视频对象分割算法  

A video object segmentation algorithm based on background extraction

作　　者：焦蓬蓬[1] 郭依正[1] 

机构地区：[1]南京师范大学泰州学院信息科学与技术系,江苏泰州225300

出　　处：《微型机与应用》2010年第19期35-36,42,共3页Microcomputer & Its Applications

摘　　要：提出了一种基于背景提取的视频对象分割算法,其主要是通过对多帧视频序列的比较,提取出视频帧的背景,以后的视频帧序列减去该背景,就得到了运动对象。实验结果表明,该算法可以很好地应用在视频会议等背景变化不大的场合,效果较好且运算量不大。This paper presents a video object segmentation algorithm based on background extraction.Its main idea is to compare the sequences of images firstly,then to extract the background of the video sequences.Finally,we can get the video object from the following video sequences minus the background.The experimental results show that,this method can be used in occasions such as videoconference,which the background is little change.Using this method can acquire a better performance with less calculation.

关 键 词：MPEG-4 视频对象 视频分割 对称差分法 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]