行人再识别研究进展综述

Summary of research on progress of person re-identification

行人再识别是计算机视觉领域的一个极具挑战但是极具研究价值的项目。由于其良好的实用性,吸引了大批研究者投入其中。文中首先简单介绍了行人再识别的概念与传统的行人检测的差别,然后介绍了行人再识别领域面临的难点,接着重点介绍了近年来在提高识别准确率方面所做的努力,尤其是特征的提取和度量学习两方面的最新研究成果,最后,文章展望了行人再识别未来的研究趋势。

Person re-identification is a challenging,but very valuable branch in computer vision. Many researchers dedicate to this area due to its practicability. At beginning, this article takes a brief introduction of concept of person re-identification, as well as the difference between person reidentification and traditional person detection. Then attention is paid to the several difficulties that person re-identification faced. After that is the introduction of efforts in improving recognition accuracy,it especially presents the latest research results of feature extraction and metric learning. Lastly,the article analyzes the research trends of person re-identification in the future.