融合时空信息的前景／阴影视频分割算法  

A Spatiotemporal Algorithm for Video Foreground and Shadow Segmentation

作　　者：褚一平[1,2] 叶修梓[1] 黄叶珏[3] 张引[1] 张三元[1] 

机构地区：[1]浙江大学计算机科学与技术学院CAD＆CG国家重点实验室,杭州310027 [2]杭州电子科技大学计算机学院,杭州310018 [3]浙江工业职业技术学院计算机系,绍兴312000

出　　处：《模式识别与人工智能》2008年第4期546-551,共6页Pattern Recognition and Artificial Intelligence

基　　金：国家自然科学基金项目(No.60473106);国家教育部博士点基金项目(No.20060335114);长江学者和创新团队发展计划项目(No.IRT0652)资助

摘　　要：视频目标分割是视频目标跟踪、统计以及识别的基础.阴影是影响目标分割准确性的重要因素,有效对阴影进行检测与消除可提高视频目标分割的质量.本文提出一种采用状态机对阴影进行建模的方法,通过阴影模型来消除阴影.算法定义背景、阴影以及前景的势函数,利用马尔可夫随机场融合视频序列的时空邻域信息,采用Gibbs 采样算法求解最大后验概率,提高视频目标分割的质量.在不同环境下对本文算法的有效性进行测试,并与其他算法进行比较,结果证明本文算法的有效性.Video segmentation is important for video object tracking, counting and recognition. Shadows are the factors that affect the accuracy of object segmentation. Efficiently detecting and removing the shadows can improve the quality of object segmentation. An algorithm for ,video-foreground and shadow segmentation is proposed in this paper. It models shadows with state machine and the shadows are removed according to the shadow models. The potential functions for the background, shadow and foreground are defined. The spatiotemporal neighboring relationships in the video sequence are constructed by using Markov random fields. Gibbs sampling algorithm is adopted to solve the MAP problem and thus the segmentation qualityis improved . The correctness of the proposed algorithm is tested under different environments and the results demonstrate the validity of the algorithm compared with other algorithms.

关 键 词：视频目标分割 阴影检测与消除 时空邻域关系 马尔可夫随机场(MRF) 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]