Person re-identification, also known as person retrieval, is to match pedestrian images observed from non-overlapping camera views based on appearance.It receives increasing attentions in video surveillance for its important applications in threat detection, human retrieval, and multi-camera tracking. It saves a lot of human labor in exhaustively searching for a person of interest from large amounts of video sequences.

CUHK03是第一个足以进行深度学习的大规模行人重识别数据集，该数据集的图像采集于香港中文大学（CUHK）校园。数据以”cuhk-03.mat”的 MAT 文件格式存储，含有 1467 个不同的人物，由 5 对摄像头采集。

近几年ICCV、CVPR、ECCV等顶级国际会议关于行人重识别(person re-identification)的论文汇总，不定期更新。
如有疏漏或不足之处，欢迎指正。