Mask wearing detection algorithm based on improved YOLOv3

Abstract

The thesis is carried out by the student of group CSAI-21f Cui Tianyuan. The topic is «Mask wearing detection algorithm based on improved YOLOv3». The thesis is submitted to earn a master’s degree in a specialty 122 «Computer Science». This degree work starts from the idea of determining whether a face is wearing a mask or not, based on the YOLOv3 algorithm, with the intention of exploring a lightweight mask wearing detection algorithm with high detection accuracy and high speed. As a result of the thesis, an algorithm is designed that greatly reduces the network complexity while guaranteeing the detection accuracy, and the detection speed reaches 125 frames/second on the experimental platform, which occupies an absolute speed advantage compared with several excellent detection algorithms. At the same time, it has good robustness in the case of luminance scale change and occlusion, and has certain adaptability to real-time detection in complex environments.