Unsupervised Spatio-Temporal Segmentation for Extracting Moving Objects in Video Sequences

作     者：李仁杰 余松煜 王向文 LI Ren-jie YU Song-yu WANG Xiang-wen (Institute of Image Communication and Information Processing,Shanghai Jiaotong University,Shanghai 200240,China)

作者机构：Institute of Image Communication and information ProcessingShanghai Jiaotong University Institute of Image Communication and Information ProcessingShanghai Jiaotong University 

出 版 物：《Journal of Shanghai Jiaotong university(Science)》 (上海交通大学学报（英文版）)

年 卷 期：2009年第14卷第2期

页      面：154-161页

核心收录：

学科分类：0711[理学-系统科学] 07[理学] 08[工学] 080203[工学-机械设计及理论] 080401[工学-精密仪器及机械] 0804[工学-仪器科学与技术] 080402[工学-测试计量技术及仪器] 0802[工学-机械工程] 

基　　金：the National Natural Science Foundationof China(No.60502034) 

主　　题：segmentation spatio-temporal region growing outlier rejection watershed 

摘      要：Extraction of moving objects is an important and fundamental research topic for many video applications. This paper addresses an unsupervised spatio-temporal segmentation scheme to extract moving objects from video *** temporal segmentation localizes moving objects by comparing the motion vector of each block in each frame with the corresponding global motion vector estimated by an outlier rejection(OR) based ***,the temporal compensation utilizing the temporal coherence of moving objects is considered in the temporal segmentation to solve the temporarily stopping *** detected moving regions usually have discontinuous boundaries and some *** regions are then compensated in the spatial domain. In the spatial segmentation,the watershed algorithm considering the global information improves the accuracy of segmentation in the spatial *** modified mean filter is presented to suppress some *** using a fusion module,moving objects are *** on various sequences have successfully demonstrated the validity of the proposed scheme.