Pedestrian attribute recognition: A survey

Pedestrian Attribute Recognition (PAR) is an important task in computer vision community and plays role practical video surveillance. The goal of this paper to review existing works using traditional methods or based on deep learning networks. Firstly, we introduce the background pedestrian attribute recognition, including fundamental concepts formulation attributes corresponding challenges. Secondly, analyze popular solutions for from eight perspectives. Thirdly, discuss specific then, give a comparison between algorithm PAR methods. After that, show connections other tasks. Fourthly, benchmark datasets, evaluation metrics community, brief performance comparison. Finally, summarize several possible research directions PAR. project page can be found at: https://sites.google.com/view/ahu-pedestrianattributes/.