A Richly Annotated Dataset for Pedestrian Attribute Recognition

Dangwei Li, Zhang Zhang, Xiaotang Chen, Haibin Ling, Kaiqi Huang


We collect a Richly Annotated Pedestrian (RAP) dataset from multi-camera surveillance scenarios for pedestrian attribute analysis. The RAP has in total 41,585 pedestrian samples, each of which is annotated with 72 attributes as well as viewpoints, occlusions, body parts information. To our knowledge, the RAP is current largest pedestrian attribute dataset, which is expected to promote the study of large-sclae attribute recognition systems.

Dataset Information


If you use RAP dataset in your research, please kindly cite our work as:


If you want to use the RAP dataset for scientific research only, please fill the aggrement here .
Recently a strong baseline has also been released.


[1] Gray, Douglas and Tao, Hai, Viewpoint invariant pedestrian recognition with an ensemble of localized features, ECCV2008
[2] Hirzer, Martin and Beleznai, Csaba and Roth, Peter M and Bischof, Horst. Person re-identification by descriptive and discriminative classification, Image Analysis 2011
[3] Liu, Chunxiao and Gong, Shaogang and Loy, Chen Change and Lin, Xinggang. Person re-identification: What features are important?, ECCV2012 Workshops
[4] Zhu, Jianqing and Liao, Shengcai and Lei, Zhen and Yi, Dong and Li, Stan. Pedestrian attribute classification in surveillance: Database and evaluation, ICCV2013 Workshops
[5] Deng, Yubin and Luo, Ping and Loy, Chen Change and Tang, Xiaoou. Pedestrian attribute recognition at far distance, Multimedia2014