Research on Mask Wearing Detection Algorithm Based on YOLOv5
Shuo Yu, Hui Li, Fangjun Gui, Yanqi Yang, Chenyang Lv · 2021 IEEE 2nd International Conference on Information Technology, Big Data and Artificial Intelligence (ICIBA) · 2021
The current mask wearing detection device is affected by numerous factors, such as the large number of people in complex scenes, the easy obstruction of the gathering crowd, and the small size of the inspection target, which are prone to false detections and missing inspections. In order to solve the above problems, this journal proposed a mask-wearing detection algorithm based on YOLOv5 to realize real-time detection in complex scenes. Firstly, perform Mosaic data enhancement and other processes on the data set. Then, apply Focus process to retain more complete down sampling information of the images for subsequent feature extraction. Later, employ SPP to integrate multi-scale information to achieve feature enhancement and retain spatial information within Neck. Finally, CIoU Loss function is selected s to improve the positioning accuracy. And during the training process, a dynamic adjustment strategy is adopted for the learning rate.Experimental results show that the average accuracy of the improved algorithm reaches 99.1%.