面向拥挤行人检测的CA-YOLOv5

doi:10.3778/j.issn.1002-8331.2201-0058

摘要/Abstract

摘要： 针对YOLOv5在拥挤行人检测任务中漏检率高、特征融合不充分等问题，提出了CA-YOLOv5行人检测算法。针对原主干网络对细粒度特征融合不充分的问题，采用Res2Block重建YOLOv5的主干网络，以提升网络的细粒度特征融合能力，提高检测精度。针对数据集目标尺度变化大的问题，引入coordinate attention（CA）模块增强感受野，增强模型对目标的精确定位能力。针对FPN结构在特征融合时导致多尺度特征表达能力下降的问题，提出特征增强模块，以增强多尺度特征的表达能力。通过结构重参数化的方法减少模型的计算量与参数量，加快目标检测速度。针对行人检测任务中普遍存在的拥挤行人问题，提出EViT模块，增强模型关注局部信息的能力，提高检测精度。实验证明，在拥挤行人检测任务中，CA-YOLOv5的检测精度达到84.86%，相较于原算法提高了3.75%，检测速度可以达到51?FPS，具有较好的检测精度与实时性。因此，CA-YOLOv5可以更好地应用于实时行人检测任务中。

关键词: 深度学习, YOLOv5, 拥挤行人检测, Res2Net

Abstract: Aiming at the problem of high miss-detection rate and insufficient feature fusion of YOLOv5 in crowded pedestrian detection task, the CA-YOLOv5 pedestrian detection algorithm is proposed. To solve the problem of insufficient fine-grained feature fusion in the original backbone network, Res2Block is used to rebuild the backbone network of YOLOv5, so as to improve the fine-grained feature fusion ability of the network and improve the detection accuracy. For the large change of target scale in dataset, coordinate attention is introduced to enhance the receptive field and the model’s ability to accurately locate the target. Aiming at the problem that FPN structure reduces the multi-scale feature expression ability during feature fusion, the feature enhancement module is proposed to enhance the multi-scale feature expression ability. Through the structural re-parameterization method to reduce the number of parameters and computation in the model, and speed up the detection. Aiming at the common problem of crowded pedestrians in pedestrian detection task, EViT is proposed to enhance the ability of the model to pay attention to local information and improve the detection accuracy. Experimental results show that in the crowded pedestrian detection task, the detection accuracy of CA-YOLOv5 reaches 84.86%, 3.75% higher than the original algorithm, and the detection speed can reach 51?FPS, which has good detection accuracy and real-time. Therefore, it can be better applied to real-time pedestrian detection task.

Key words: deep learning, YOLOv5, crowded pedestrian detection, Res2Net

陈一潇, 阿里甫·库尔班, 林文龙, 袁旭. 面向拥挤行人检测的CA-YOLOv5[J]. 计算机工程与应用, 2022, 58(9): 238-245.

CHEN Yixiao, Alifu·Kuerban, LIN Wenlong, YUAN Xu. CA-YOLOv5 for Crowded Pedestrian Detection[J]. Computer Engineering and Applications, 2022, 58(9): 238-245.

参考文献

[1] 童玉芬.中国人口的最新动态与趋势——结合第七次全国人口普查数据的分析[J].中国劳动关系学院学报，2021，35（4）：15-25.
TONG Y F.Research on the latest dynamics and trends of population in China—an analysis combined with the seventh national census data[J].Journal of China University of Labor Relations，2021，35（4）：15-25.
[2] 任梦茹，侯宏录，韩修来.Gabor特征结合快速HOG特征的行人检测[J].计算机系统应用，2021，30（10）：259-263.
REN?M R，HOU?H L，HAN?X L.Pedestrian detection based on Gabor feature combined with fast HOG feature[J].Computer?Systems?&?Applications，2021，30（10）：259-263.
[3] 溪海燕，肖志涛，张芳.基于线性SVM的车辆前方行人检测方法[J].天津工业大学学报，2012，31（1）：69-73.
XI H Y，XIAO Z T，ZHANG F.Method of pedestrian detection ahead of vehicle based on linear SVM[J].Journal of Tianjin Polytechnic University，2012，31（1）：69-73.
[4] TANG S，ANDRILUKA M，SCHIELE B.Detection and tracking of occluded people[J].International Journal of Computer Vision，2014，110（1）：58-69.
[5] GIRSHICK R，DONAHUE J，DARRELL T，et al.Rich feature hierarchies for accurate object detection and semantic segmentation[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition，2014：580-587.
[6] GIRSHICK R.Fast R-CNN[C]//Proceedings of the IEEE International Conference on Computer Vision，2015：1440-1448.
[7] REN S，HE K，GIRSHICK R，et al.Faster R-CNN：towards real-time object detection with region proposal networks[J].Advances in Neural Information Processing Systems，2015，28：91-99.
[8] HE K，ZHANG X，REN S，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.
[9] REDMON J，DIVVALA S，GIRSHICK R，et al.You only look once：unified，real-time object detection[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition，2016：779-788.
[10] REDMON J，FARHADI A.YOLO9000：better，faster，stronger[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition，2017：7263-7271.
[11] REDMON J，FARHADI A.Yolov3：an incremental improvement[J].arXiv：1804.02767，2018.
[12] BOCHKOVSKIY A，WANG C Y，LIAO H Y M.YOLOv4：Optimal speed and accuracy of object detection[J].arXiv：2004.10934，2020.
[13] LIU W，ANGUELOV D，ERHAN D，et al.SSD：single shot multibox detector[C]//European Conference on Computer Vision.Cham：Springer，2016：21-37.
[14] TIAN Z，SHEN C，CHEN H，et al.Fcos：fully convolutional one-stage object detection[C]//Proceedings of the IEEE/CVF International Conference on Computer Vision，2019：9627-9636.
[15] 张立旗，肖秦琨，韩泽佳.基于改进YOLOv3的行人检测算法[J].自动化与仪表，2021，36（10）：57-62.
ZHANG L Q，XIAO Q K，HAN Z J.Pedestrian detection algorithm based on CA YOLOv3[J].Automation & Instrumentation，2021，36（10）：57-62.
[16] 胡均平，孙希.基于改进Faster R-CNN的近红外夜间行人检测方法[J].传感器与微系统，2021，40（8）：126-129.
HU J P，SUN X.Near IR night pedestrian detection method based on CA faster R-CNN[J].Transducer and Microsystem Technologies，2021，40（8）：126-129.
[17] 舒壮壮，单梁，马苗苗，等.基于YOLOv3的改进行人检测算法研究[J].南京理工大学学报，2021，45（3）：259-264.
SHU Z Z，SHAN M，MA M M，et al.CA pedestrian detection algorithm based on YOLOv3[J].Journal of Nanjing University of Science and Technology，2021，45（3）：259-264.
[18] HE K，ZHANG X，REN S，et al.Deep residual learning for image recognition[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition，2016：770-778.
[19] HUANG G，LIU Z，VAN DER MAATEN L，et al.Densely connected convolutional networks[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition，2017：4700-4708.
[20] TAN M，LE Q.Efficientnet：rethinking model scaling for convolutional neural networks[C]//International Conference on Machine Learning，2019：6105-6114.
[21] LIU S，QI L，QIN H，et al.Path aggregation network for instance segmentation[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition，2018：8759-8768.
[22] WANG C Y，LIAO H Y M，WU Y H，et al.CSPNet：a new backbone that can enhance learning capability of CNN[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops，2020：390-391.
[23] GAO S，CHENG M M，ZHAO K，et al.Res2net：a new multi-scale backbone architecture[J].IEEE Transactions on Pattern Analysis and Machine Intelligence，2021，43（2）：652-662.
[24] HOU Q，ZHOU D，FENG J.Coordinate attention for efficient mobile network design[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition，2021：13713-13722.
[25] GUO C，FAN B，ZHANG Q，et al.Augfpn：improving multi-scale feature learning for object detection[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition，2020：12595-12604.
[26] YU F，KOLTUN V.Multi-scale context aggregation by dilated convolutions[J].arXiv：1511.07122，2015.
[27] DING X，ZHANG X，MA N，et al.Repvgg：making vgg-style convnets great again[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition，2021：13733-13742.
[28] DOSOVITSKIY A，BEYER L，KOLESNIKOV A，et al.An image is worth 16×16 words：Transformers for image recognition at scale[J].arXiv：2010.11929，2020.
[29] SHAO S，ZHAO Z，LI B，et al.Crowdhuman：a benchmark for detecting human in a crowd[J].arXiv：1805.00123，2018.
[30] LI J，WANG C，ZHU H，et al.Crowdpose：efficient crowded scenes pose estimation and a new benchmark[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition，2019：10863-10872.