摘要
人体目标分割是人体目标视觉分析的关键问题之一。提出了一种基于Mean-shift的粘连人体目标分割算法。首先对视频图像进行预处理,从中分离出运动区域,根据人体外形的统计特征建立人体目标模板。在运动区域中均匀取若干个数据点作为种子点。从种子点出发,基于人体目标模板,应用Mean-shift算法不断迭代逼近模态点。对取得的模态点集合进行聚类,从而自动确定分类数,即运动区域中的人体目标数,并进行合理分割。基于PETS2006数据库的试验验证了该方法的可行性。
Human objects Segmentation is one of the key problems of Visual Analysis. In this paper, a novel touched human objects segmentation based on Mean-shift algorithm was proposed. At first, Video Images was preprocessed and motion regions were obtained, and model of human object was built according to statistical characteristics of body surface. Then, a few of points of motion region picked equably were taken as seeds, and local mode centroids were calculated by Mean-shift iterative process. At last, the number of categories was automatically acqured based on the clustering algorithm, and human objects were segmented according to the result of clustering. The experiment based on PETS 2006 Database proves this method is feasible.
出处
《计算机应用》
CSCD
北大核心
2009年第1期51-53,共3页
journal of Computer Applications
基金
深圳市科技计划项目(SZKJ0702)
广东省自然科学基金资助项目(7008732)