Occlusion and multi-scale pedestrian detection A review

Abstract

Pedestrian detection has a wide range of application prospects in many fields such as unmanned driving, intelligent monitoring, robot, etc., and has always been a hot issue in the field of computer vision. In recent years, with the development of deep learning and the proposal of many large pedestrian data sets, pedestrian detection technology has also made great progress, and the detection accuracy and detection speed have been significantly improved. However, the performance of the most advanced pedestrian detection methods is still far behind that of human beings, especially when there is occlusion and scale change, the detection accuracy decreases significantly. Occlusion and scale problems are the key problems to be solved in pedestrian detection. The purpose of this paper is to discuss the research progress of pedestrian detection. Firstly, this paper explores the research status of pedestrian detection in the past four years (2019–2022), focuses on analyzing the occlusion and scale problems of pedestrian detection and corresponding solutions, summarizes the data sets and evaluation methods of pedestrian detection, and finally looks forward to the development trend of the occlusion and scale problems of pedestrian detection.