Pedestrian Attribute Recognition Based On Deep Learning : A Survey

With the continuous development of intelligent video surveillance, the research on pedestrian attributes has attracted widespread attention. The paper aims to introduce the existing pedestrian attribute recognition based on deep learning. We summarized the existing work on pedestrian attribute recognition. The research of deep learning on pedestrian attribute recognition in terms global-based, locally-based, of GCN-based and attention mechanisms. At the same time, we introduced the related background and concepts of pedestrian attribute recognition. Then, we introduced its commonly used datasets and evaluation criteria. At the end, we summarize the current research hotspots and future research directions.