摘 要：新型冠状病毒传播方式主要为飞沫传播和近距离接触传播等，而正确佩戴口罩则是预防、隔离病毒的最方便有效的手段之一。但在公共场合下仍存在部分人群未佩戴口罩的情况，给公共安全带来了巨大威胁。从 WIDER Face、MAFA、RMFD 和 MaskedFace-Net 四个公开数据集中筛选 7 240 张图像，构成人脸口罩规范数据集用于算法训练与测试。结果表明，YOLOv4 算法在检测精度和检测速度方面还不错，检测效果满足场景需求。基于 YOLOv4 的口罩规范检测算法可以有效减轻防疫人员的工作量，提高了人员密集公共场合的安全系数。
关键词：新冠疫情；口罩检测；深度学习；YOLOv4
DOI:10.19850/j.cnki.2096-4706.2022.21.007
中图分类号：TP391.4 文献标识码：A 文章编号：2096-4706（2022）21-0029-04
Research on Mask Specification Detection Algorithm Based on YOLOv4
LU Litian
(School of Computer and Information Science, Chongqing Normal University, Chongqing 401331, China)
Abstract: The transmission methods of the new coronavirus are mainly droplet transmission and close contact transmission, and wearing a mask correctly is one of the most convenient and effective means to prevent and isolate the virus. However, some people still do not wear masks in public places, posing a huge threat to public safety. 7240 images are selected from four public datasets of WIDER Face, MAFA, RMFD and MaskedFace-Net to form a face mask specification dataset for algorithm training and testing. The results show that the YOLOv4 algorithm is good in terms of detection accuracy and detection speed, and the detection effect meets the needs of the scene. The mask specification detection algorithm based on YOLOv4 can effectively reduce the workload of epidemic prevention personnel and improve the safety factor in crowded public places.
Keywords: covid-19; mask detection; deep learning; YOLOv4
参考文献：
[1] CHAVEZ S，LONG B，KOYFMAN A，et al. Coronavirus Disease （COVID-19）：A primer for emergency physicians [J].The American journal of emergency medicine，2021，44：220-229.
[2] REDMON J，DIVVALA S，GIRSHICK R，et al. You Only look Once：Unified， Real-time Object Detection [J/OL]. arXiv:1506.02640 [cs.CV].[2022-07-03].https://arxiv.org/abs/1506.02640.
[3] SZEGEDY C，LIU W，JIA Y Q，et al. Going deeper with convolutions [J/OL].arXiv：1409.4842 [cs.CV].[2022-07-03].https://arxiv. org/abs/1409.4842.
[4] REDMON J，FARHADI A. YOLO9000：better，faster， stronger [C]//2017 IEEE Conference on Computer Vision and Pattern Recognition（CVPR）.Honolulu：IEEE，2017：6517-6525.
[5] IOFFE S，SZEGEDY C. Batch normalization：Accelerating deep network training by reducing internal covariate shift [J/OL].[2022- 07-03].https://www.doc88.com/p-9999653650220.html.
[6] REN S Q，HE K M，GIRSHICK R，et al. Faster R-Cnn： Towards Real-Time Object Detection with Region Proposal Networks [J].IEEE Transactions on Pattern Analysis and Machine Intelligence， 2017，39（6）：1137-1149.
[7] REDMON J，FARHADI A. Yolov3：An incremental improvement [J/OL].arXiv：1804.02767 [cs.CV].[2022-07-03].https:// arxiv.org/abs/1804.02767.
[8] HE K M，ZHANG X Y，REN S Q，et al. Deep Residual Learning for Image Recognition [C]//2016 IEEE Conference on Computer Vision and Pattern Recognition（CVPR）.Las Vegas： IEEE，2016：770-778.
[9] LIN T Y，DOLLAR P，GIRSHICK R，et al. Feature Pyramid Networks for Object Detection [J/OL].arXiv：1612.03144 [cs.CV].[2022- 07-03].https://arxiv.org/abs/1612.03144.
[10] BOCHKOVSKIY A，WANG C Y，LIAO H Y M. Yolov4： Optimal speed and accuracy of object detection [J/OL].arXiv： 2004.10934 [cs.CV].[2022-07-03].https://arxiv.org/abs/2004.10934.
[11] WANG C Y，LIAO H Y M，WU Y H，et al. CSPNet： A new backbone that can enhance learning capability of CNN [C]//2020 IEEE/CVF conference on computer vision and pattern recognition workshops CVPRW.Seattle：IEEE，2020：390-391.
[12] LIU S，QI L，QIN H F，et al. Path Aggregation Network for Instance Segmentation [C]//2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition.Salt Lake City：IEEE，2018：8759-8768.
[13] HE K M，ZHANG X Y，REN S Q，et al. Spatial pyramid pooling in deep convolutional networks for visual recognition [J].IEEE transactions on pattern analysis and machine intelligence，2015，37（9）： 1904-1916.
[14] HARTIGAN J A，WONG M A. Algorithm AS 136：A k-means clustering Algorithm [J].Journal of the royal statistical society. series c （applied statistics），1979，28（1）：100-108.
[15] HOSANG J，BENENSON R，SCHIELE B. Learning nonmaximum suppression [C]//2017 IEEE conference on computer vision and pattern recognition.Honolulu：IEEE，2017：4507-4515.
[16] ZHENG Z H，WANG P，LIU W，et al. Distance-IoU loss： Faster and better learning for bounding box regression [J].Proceedings of the AAAI Conference on Artificial Intelligence，2020，34（7）：12993-13000.
[17] JIANG B R，LUO R X，MAO J Y，et al．Acquisition of Localization Confidence for Accurate Object Detection ［J/OL］.arXiv： 1807.11590 [cs.CV].[2022-07-03].https://arxiv.org/abs/1807.11590v1.
[18] YANG S，LUO P，LOY C C，et al. Wider face：A face detection benchmark [C]//2016 IEEE Conference on Computer Vision and Pattern Recognition CVPR.Las Vegas：IEEE，2016：5525-5533.
[19] WANG Z Y，WANG G C，HUANG B J，et al. Masked face recognition dataset and application [J/OL].arXiv：2003.09093 [cs.CV]. [2022-07-03].https://arxiv.org/abs/2003.09093v2.
[20] GE S M，LI J，YE Q T，et al. Detecting masked faces in the wild with LLE-CNNs [C]//2017 IEEE Conference on Computer Vision and Pattern（CVPR）.Honolulu：IEEE，2017：426-434.
[21] CABANI A，HAMMOUDI K，BENHABILES H，et al. “Maskedface-Net–A dataset of correctly/incorrectly masked face images in the context of COVID-19 [J].Smart Health，2021，19：100144.
作者简介：陆立天（1996—），男，汉族，湖北武汉人，硕士研究生在读，研究方向：深度学习与计算机视觉。