Application of Improved YOLOv3 Algorithm in Mask Recognition

摘要

With the influence of novel coronavirus, wearing masks is becoming more and more important. If computer vision system is used in public places to detect whether a pedestrian is wearing a mask, it will improve the efficiency of social operation. Therefore, a new mask recognition algorithm based on improved yolov3 is proposed. Firstly, the dataset is acquired through network video; secondly, the dataset is pre-processed; finally, a new network model is proposed and the activation function of YOLOv3 is changed. The average accuracy of the improved YOLOv3 algorithm is 83.79%. This method is 1.18% higher than the original YOLOv3.