1.青岛大学 自动化学院, 山东 青岛 266071
2.山东省工业控制技术重点实验室, 山东 青岛 266071
[ "付惠琛(1999—),男,山东烟台人,硕士研究生,2021年于青岛大学获得学士学位。主要从事计算机视觉与模式识别方面的研究。E-mail:fuhuichen1999@163.com" ]
[ "高军伟(1972—),男,山东临沂人,博士,教授,2003年于中国铁道科学研究院获得博士学位。主要从事模式识别及智能控制方面的研究。E-mail:qdgao163@163.com" ]
扫 描 看 全 文
付惠琛, 高军伟, 车鲁阳. 基于改进YOLOv7的口罩佩戴检测[J]. 液晶与显示, 2023,38(8):1139-1147.
FU Hui-chen, GAO Jun-wei, CHE Lu-yang. Mask wearing detection based on improved YOLOv7[J]. Chinese Journal of Liquid Crystals and Displays, 2023,38(8):1139-1147.
付惠琛, 高军伟, 车鲁阳. 基于改进YOLOv7的口罩佩戴检测[J]. 液晶与显示, 2023,38(8):1139-1147. DOI: 10.37188/CJLCD.2022-0371.
FU Hui-chen, GAO Jun-wei, CHE Lu-yang. Mask wearing detection based on improved YOLOv7[J]. Chinese Journal of Liquid Crystals and Displays, 2023,38(8):1139-1147. DOI: 10.37188/CJLCD.2022-0371.
佩戴好口罩是居民预防新冠和配合国家疫情防控的有效方式。针对口罩佩戴是否正确、拍摄角度不同以及被遮挡等问题,提出了一种改进的YOLOv7算法。该算法以YOLOv7为基础,在网络的Head区引入卷积注意力机制,使得特征网络在对口罩区域的处理中更具有针对性,从而增强特征网络对口罩区域的学习能力;对Backbone区结构进行优化,对ConvNeXt网络结构进行改进,并引入网络中代替部分卷积,提高模型的检测精度和鲁棒性,增强预测精确度的同时不会引入大量额外的计算。对Head层的空间金字塔池化进行改进,提高了训练速度并且加快模型收敛。实验结果表明,在复杂及遮挡的情况下,改进后的YOLOv7的损失函数大幅下降,在测试集上的mAP为93.8%,相比于原始YOLOv7算法提高了3.6%。各个类别的检测精度均有提升,没佩戴口罩、正确佩戴口罩、不正确佩戴口罩类别的精度分别提升6.8%、2.1%、1.7%。本文算法的错检情况明显减少,泛化能力有显著提升。
Wearing masks is an effective way for preventing COVID-19 and cooperating with the national epidemic prevention and control. An improved YOLOv7 algorithm is proposed to solve the problems such as whether masks are correctly worn, different shooting angles and being blocked. Based on YOLOv7, the convolutional attention mechanism is introduced into the Head region of the network to make the feature network more targeted in the processing of the mask region, thus enhancing the learning ability of the feature network to the mask region. The structure of Backbone area is optimized, the ConvNeXt network structure is improved, and partial convolution is introduced into the network instead, which improves the detection accuracy and robustness of the model and enhances the accuracy of prediction without introducing a large number of additional calculations. The space pyramid pool of the Head layer is improved to improve the training speed and accelerate the model convergence. Experiments show that in the case of complexity and occlusion, the loss function of the improved YOLOv7 decreases significantly, and the mAP on the test set is 93.8%, which is 3.6% higher than that of the original YOLOv7 algorithm.The accuracy of each category is improved, and the accuracy of no mask, correct mask and incorrect mask are increased by 6.8%, 2.1% and 1.7%, respectively. The cases of error detection are significantly reduced, and the generalization ability is significantly improved.
图像处理目标检测YOLOv7算法卷积注意力机制空间金字塔池化
image processingobject detectionYOLOv7 algorithmconvolutional attention mechanismspace pyramid pooling
马丝妮,包刚升.“平衡抗疫”:前奥密克戎时期的新冠疫情防控研究[J].学术月刊,2022,54(4):78-99.
MA S N, BAO G S. “Balanced anti-epidemic”: a study on the prevention and control of the COVID-19 epidemic in the pre-omicron period [J]. Academic Monthly, 2022, 54(4): 78-99. (in Chinese)
曹素珍,温东森,陈星,等.新冠肺炎疫情期间我国居民佩戴口罩防护行为研究[J].环境科学研究,2020,33(7):1649-1658.
CAO S Z, WEN D S, CHEN X, et al. Protective behavior of Chinese population wearing masks during the COVID-19 epidemic [J]. Research of Environmental Sciences, 2020, 33(7): 1649-1658. (in Chinese)
SITU G H. Deep holography [J]. Light: Advanced Manufacturing, 2022, 3(2): 278-300. doi: 10.37188/lam.2022.013http://dx.doi.org/10.37188/lam.2022.013
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. doi: 10.1109/tpami.2016.2577031http://dx.doi.org/10.1109/tpami.2016.2577031
REDMON J, FARHADI A. YOLO9000: better, faster, stronger [C]//Proceedings of 2017 IEEE Conference on Computer Vision and Pattern Recognition. Honolulu: IEEE, 2017: 6517-6525. doi: 10.1109/cvpr.2017.690http://dx.doi.org/10.1109/cvpr.2017.690
曾广华,杨桂忠,郭寿南,等.口罩实时检测系统的设计与应用[J].电视技术,2022,46(9):65-67.
ZENG G H, YANG G Z, GUO S N, et al. Design and application of real-time mask detection system [J]. Video Engineering, 2022, 46(9): 65-67. (in Chinese)
朱杰,王建立,王斌.基于YOLOv4-tiny改进的轻量级口罩检测算法[J].液晶与显示,2021,36(11):1525-1534. doi: 10.37188/CJLCD.2021-0059http://dx.doi.org/10.37188/CJLCD.2021-0059
ZHU J, WANG J L, WANG B. Lightweight mask detection algorithm based on improved YOLOv4-tiny [J]. Chinese Journal of Liquid Crystals and Displays, 2021, 36(11): 1525-1534. (in Chinese). doi: 10.37188/CJLCD.2021-0059http://dx.doi.org/10.37188/CJLCD.2021-0059
郑欣,田博,李晶晶.基于YOLO模型的宫颈细胞簇团智能识别方法[J].液晶与显示,2018,33(11):965-971. doi: 10.3788/yjyxs20183311.0965http://dx.doi.org/10.3788/yjyxs20183311.0965
ZHENG X, TIAN B, LI J J. Intelligent recognition method of cervical cell cluster based on YOLO model [J]. Chinese Journal of Liquid Crystals and Displays, 2018, 33(11): 965-971. (in Chinese). doi: 10.3788/yjyxs20183311.0965http://dx.doi.org/10.3788/yjyxs20183311.0965
李国友,李晨光,王维江,等.基于单样本学习的多特征人体姿态模型识别研究[J].光电工程,2021,48(2):200099. doi: 10.12086/oee.2021.200099http://dx.doi.org/10.12086/oee.2021.200099
LI G Y, LI C G, WANG W J, et al. Research on multi-feature human pose model recognition based on one-shot learning [J]. Opto-electronic Engineering, 2021, 48(2): 200099. (in Chinese). doi: 10.12086/oee.2021.200099http://dx.doi.org/10.12086/oee.2021.200099
马双双,王佳,曹少中,等.基于深度学习的二维人体姿态估计算法综述[J].计算机系统应用,2022,31(10):36-43.
MA S S, WANG J, CAO S Z, et al. Overview on two-dimensional human pose estimation methods based on deep learning [J]. Computer Systems & Applications, 2022, 31(10): 36-43. (in Chinese)
LUO Y, ZHAO Y F, LI J X, et al. Computational imaging without a computer: seeing through random diffusers at the speed of light [J]. eLight, 2022, 2: 4. doi: 10.1186/s43593-022-00012-4http://dx.doi.org/10.1186/s43593-022-00012-4
张润梅,毕利君,汪方斌,等.多尺度特征融合与锚框自适应的目标检测算法[J].激光与光电子学进展,2022,59(12):1215019. doi: 10.3788/LOP202259.1215019http://dx.doi.org/10.3788/LOP202259.1215019
ZHANG R M, BI L J, WANG F B, et al. Multiscale feature fusion and anchor adaptive object detection algorithm [J]. Laser & Optoelectronics Progress, 2022, 59(12): 1215019. (in Chinese). doi: 10.3788/LOP202259.1215019http://dx.doi.org/10.3788/LOP202259.1215019
丁勇,王翔,严晓浪.边缘自适应的四点分段抛物线图像缩放[J].浙江大学学报(工学版),2010,44(9):1637-1642.
DING Y, WANG X, YAN X L. Edge adaptive four-point piecewise parabolic scaler implementation [J]. Journal of Zhejiang University (Engineering Science), 2010, 44(9): 1637-1642. (in Chinese)
HU C P, BAI X, QI L, et al. Vehicle color recognition with spatial pyramid deep learning [J]. IEEE Transactions on Intelligent Transportation Systems, 2015, 16(5): 2925-2934. doi: 10.1109/tits.2015.2430892http://dx.doi.org/10.1109/tits.2015.2430892
ZUO C, QIAN J M, FENG S J, et al. Deep learning in optical metrology: a review [J]. Light: Science & Applications, 2022, 11(1): 39. doi: 10.1038/s41377-022-00714-xhttp://dx.doi.org/10.1038/s41377-022-00714-x
FENG Y B, YANG X, QIU D W, et al. PCXRNet: pneumonia diagnosis from chest X-ray images using condense attention block and multiconvolution attention block [J]. IEEE Journal of Biomedical and Health Informatics, 2022, 26(4): 1484-1495. doi: 10.1109/jbhi.2022.3148317http://dx.doi.org/10.1109/jbhi.2022.3148317
DOSOVITSKIY A, BEYER L, KOLESNIKOV A, et al. An image is worth 16×16 words: transformers for image recognition at scale [C]. 9th International Conference on Learning Representations. Seattle: OpenReview.net, 2021: 1909-1931.
YANG X K, ZHAO J Y, ZHANG H Y, et al. Remote sensing image detection based on YOLOv4 improvements [J]. IEEE Access, 2022, 10: 95527-95538. doi: 10.1109/access.2022.3204053http://dx.doi.org/10.1109/access.2022.3204053
TANG Y L, GONG W G, CHEN X, et al. Deep inception-residual Laplacian pyramid networks for accurate single-image super-resolution [J]. IEEE Transactions on Neural Networks and Learning Systems, 2020, 31(5):1514-1528. doi: 10.1109/tnnls.2019.2920852http://dx.doi.org/10.1109/tnnls.2019.2920852
0
浏览量
80
下载量
0
CSCD
关联资源
相关文章
相关作者
相关机构