基于特征表示的行人再识别技术综述

Survey of pedestrian re-identification with feature representation

行人再识别是智能视频监控中的一项关键任务,是近年来计算机视觉领域中一直备受关注的研究热点,适用于安防以及公共场所寻人等技术领域。特征提取是行人再识别技术中存在的核心问题之一。对现有的基于特征表示的行人再识别方法进行评述,并分析其中代表性方法的优缺点;介绍了常用行人数据库的特点;然后总结现阶段行人再识别研究所面临的挑战。最后对行人再识别技术的未来发展方向进行了展望。

Pedestrian re-identification is a key task in intelligent video surveillance and a major concernin the field ol computer vision in recent years. A nd the technique can be applied in different importantapplications,e. g. security and finding someone in the public place. Feature extraction is one of the coreissues of pedestrian re-identification technique,this paper reviews the existing methods based on featurerepresentation. According to the research,the advantages and disadvantages of different typical methodsare discussed; it describes the characteristics of commonly used pedestrian re-identification database;then,it summarizes the challenges of pedestrian re-identification technique faced by the institute at thisstage. Finally,the development direction of pedestrian re-identificationin the future is proposed.