Current researches and trends of crowd counting in the field of deep learning

Crowd counting is a branch of computer vision, which has important practical significance. This paper summarizes the current situation and development trend of crowd counting based on computer vision. Firstly introducing the methods of crowd counting with shallow learning, and then presenting the research status and trend of deep learning in crowd counting. Then, the performance evaluation of crowd counting method is introduced, and several benchmark datasets for crowd counting test and assessment are provided. Finally, the focus of research at the current stage is affirmed, and the prospect is made.