基于改进YOLOv5的多任务安全人头检测算法

doi:10.19678/j.issn.1000-3428.0063232

摘要/Abstract

摘要： 针对目前监控摄像头预置检测算法存在人脸隐私泄露风险的问题，提出一种基于YOLOv5结构的多任务安全人头检测算法Privacyface-YOLO及轻量级检测版本Privacyface-YOLO （Light）。Privacyface-YOLO的主干网络采用带残差结构的跨阶段局部网络ResCSP，通过残差结构降低特征学习的复杂度，提升网络对人头特征的提取能力。轻量级版本Privacyface-YOLO （Light）使用深度可分离卷积替换ResCSP模块，以减少计算量并提升网络的运行速度。模型的颈部模块引入特征金字塔与路径聚合网络，实现跨层多尺度信息交流，提升模型在复杂人头与小目标场景下的鲁棒性。对提取出的特征进行三分支输出，其中目标定位分支用于定位人头坐标，人头分割分支生成覆盖人头的掩膜以保证人脸隐私，人数回归分支获得图像中的人头总数，通过多任务三分支结构保护人脸隐私同时满足更加复杂的场景需求。实验结果表明，Privacyface-YOLO算法能够有效完成实时人头检测任务并保护人脸隐私，相较目前主流人头检测算法，在人头数据集HollywoodHeads和Brainwash上，该算法的平均查准率AP50指标分别提升11.8%和5.8%，平均查准率AP70指标分别提升20.2%和35.2%。

关键词: 人头检测, 深度学习, YOLOv5算法, 人脸隐私, 目标检测

Abstract: Considering the risk of privacy violations that could leak facial information in existing detection algorithms used by surveillance cameras, this study proposes a multi-task secure head-detection algorithm based on a YOLOv5 structure, called Privacyface-YOLO, as well as a lightweight version called Privacyface-YOLO(Light).The backbone network of Privacyface-YOLO adopts the cross-stage local network ResCSP with a residual structure, which reduces the complexity of feature learning and improves the ability of the network to extract features.The lightweight version Privacyface-YOLO(Light) replaces the ResCSP module with deep separable convolution to reduce the amount of computation performed and improve running speed.A neck module introduces a feature pyramid and a path aggregation network to realize cross-layer and multi-scale information exchange and to improve robustness to scenes with small targets and complex head shapes.The extracted features are output in three branches that perform different tasks to protect privacy and meet the needs of complex scenes.A target-location branch is used to locate the coordinates of people's heads in an image, and a head-segmentation branch generates a mask covering each head to ensure the privacy of face information, while a dedicated regression branch obtains the total number of heads in the image.Experimental results are presented to show that the Privacyface-YOLO algorithm can perform the real-time head-detection task effectively while protecting the privacy of face information.The proposed algorithm outperformed existing mainstream head-detection algorithms, obtaining a higher AP50 average precision index on the HollywoodHeads and Brainwash datasets by 11.8% and 5.8%, respectively;its AP70 average precision index was higher by 20.2% and 35.2%, respectively.

Key words: head-detection, deep learning, YOLOv5 algorithm, face privacy, target detection

中图分类号:

TP391

毛雨晴, 赵奎. 基于改进YOLOv5的多任务安全人头检测算法[J]. 计算机工程, 2022, 48(8): 136-143.

MAO Yuqing, ZHAO Kui. Multi-Task Secure Head-Detection Algorithm Based on Improved YOLOv5[J]. Computer Engineering, 2022, 48(8): 136-143.

http://www.ecice06.com/CN/Y2022/V48/I8/136

图/表 10

20220825091904

20220825091908

20220825091913

20220825091919

20220825091922

20220825091926

20220825091930

20220825091933

20220825091937

20220825091940

参考文献

[1] 方禹.论对《网络安全法》第67条中"个人信息"的理解[J].网络信息法学研究, 2020(1):15. FANG Y.A discussion of the definition of "personal information" in the network security law[J].Journal of Cyber and Information Law, 2020(1):15.(in Chinese)
[2] VIOLA P, JONES M J.Robust real-time face detection[J].International Journal of Computer Vision, 2004, 57(2):137-154.
[3] CHEN J, SHAN S G, YANG S Y, et al.Modification of the AdaBoost-based detector for partially occluded faces[C]//Proceedings of the 18th International Conference on Pattern Recognition.Washington D.C., USA:IEEE Press, 2006:516-519.
[4] LOWE D G.Distinctive image features from scale-invariant keypoints[J].International Journal of Computer Vision, 2004, 60(2):91-110.
[5] DALAL N, TRIGGS B.Histograms of oriented gradients for human detection[EB/OL].[2021-10-05].http://vision.stanford.edu/teaching/cs231b_spring1213/papers/CVPR05_DalalTriggs.pdf.
[6] YAN S Y, SHAN S G, CHEN X L, et al.Locally Assembled Binary(LAB) feature with feature-centric cascade for fast and accurate face detection[C]//Proceedings of 2008 IEEE Conference on Computer Vision and Pattern Recognition.Washington D.C., USA:IEEE Press, 2008:1-7.
[7] BAY H, ESS A, TUYTELAARS T, et al.Speeded-Up Robust Features(SURF)[J].Computer Vision and Image Understanding, 2008, 110(3):346-359.
[8] WU S Z, KAN M N, HE Z L, et al.Funnel-structured cascade for multi-view face detection with alignment-awareness[J].Neurocomputing, 2017, 221:138-145.
[9] VU T H, OSOKIN A, LAPTEV I.Context-aware CNNs for person head detection[C]//Proceedings of 2015 IEEE International Conference on Computer Vision.Washington D.C., USA:IEEE Press, 2015:2893-2901.
[10] 吉训生, 吴凡.基于混合卷积神经网络的人头检测方法[J].高技术通讯, 2018, 28(4):313-319. JI X S, WU F.Head detection using hybrid convolution neural networks[J].Chinese High Technology Letters, 2018, 28(4):313-319.(in Chinese)
[11] LI W, LI H L, WU Q B, et al.HeadNet:an end-to-end adaptive relational network for head detection[J].IEEE Transactions on Circuits and Systems for Video Technology, 2020, 30(2):482-494.
[12] WANG A, CAO X H, LU L, et al.Design of efficient human head statistics system in the large-angle overlooking scene[J].Electronics, 2021, 10(15):1851.
[13] STEWART R, ANDRILUKA M, NG A Y.End-to-end people detection in crowded scenes[C]//Proceedings of 2016 IEEE Conference on Computer Vision and Pattern Recognition.Washington D.C., USA:IEEE Press, 2016:2325-2333.
[14] VORA A, CHILAKA V.FCHD:fast and accurate head detection in crowded scenes[EB/OL].[2021-10-05].https://arxiv.org/pdf/1809.08766.pdf.
[15] LIU S, QI L, QIN H F, et al.Path aggregation network for instance segmentation[C]//Proceedings of 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition.Washington D.C., USA:IEEE Press, 2018:8759-8768.
[16] SANDLER M, HOWARD A, ZHU M L, et al.MobileNetV2:inverted residuals and linear bottlenecks[C]//Proceedings of 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition.Washington D.C., USA:IEEE Press, 2018:4510-4520.
[17] GIRSHICK R, DONAHUE J, DARRELL T, et al.Rich feature hierarchies for accurate object detection and semantic segmentation[C]//Proceedings of 2014 IEEE Conference on Computer Vision and Pattern Recognition.Washington D.C., USA:IEEE Press, 2014:580-587.
[18] LIU W, ANGUELOV D, ERHAN D, et al.SSD:single shot MultiBox detector[EB/OL].[2021-10-05].https://arxiv.org/pdf/1512.02325v2.pdf.
[19] REDMON J, DIVVALA S, GIRSHICK R, et al.You only look once:unified, real-time object detection[C]//Proceedings of 2016 IEEE Conference on Computer Vision and Pattern Recognition.Washington D.C., USA:IEEE Press, 2016:779-788.
[20] LIU L, OUYANG W L, WANG X G, et al.Deep learning for generic object detection:a survey[J].International Journal of Computer Vision, 2020, 128(2):261-318.
[21] REDMON J, FARHADI A.YOLO9000:better, faster, stronger[C]//Proceedings of 2017 IEEE Conference on Computer Vision and Pattern Recognition.Washington D.C., USA:IEEE Press, 2017:7263-7271.
[22] REDMON J, FARHADI A.YOLOv3:an incremental improvement[EB/OL].[2021-10-05].https://arxiv.org/abs/1804.02767.
[23] BOCHKOVSKIY A, WANG C Y, LIAO H Y M.YOLOv4:optimal speed and accuracy of object detection[EB/OL].[2021-10-05].https://arxiv.org/abs/2004.10934.
[24] WANG C Y, MARK LIAO H Y, WU Y H, et al.CSPNet:a new backbone that can enhance learning capability of CNN[C]//Proceedings of 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition.Washington D.C., USA:IEEE Press, 2020:1571-1580.
[25] LIN T Y, DOLLAR P, GIRSHICK R, et al.Feature pyramid networks for object detection[C]//Proceedings of 2017 IEEE Conference on Computer Vision and Pattern Recognition.Washington D.C., USA:IEEE Press, 2017:936-944.
[26] REZATOFIGHI H, TSOI N, GWAK J, et al.Generalized intersection over union:a metric and a loss for bounding box regression[C]//Proceedings of 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition.Washington D.C., USA:IEEE Press, 2019:658-666.
[27] YANG S, LUO P, LOY C C, et al.WIDER FACE:a face detection benchmark[C]//Proceedings of 2016 IEEE Conference on Computer Vision and Pattern Recognition.Washington D.C., USA:IEEE Press, 2016:5525-5533.
[28] RAHMAN M A, WANG Y.Optimizing intersection-over-union in deep neural networks for image segmentation[EB/OL].[2021-10-05].https://www.cs.umanitoba.ca/~ywang/papers/isvc16.pdf.
[29] HU P Y, RAMANAN D.Finding tiny faces[C]//Proceedings of 2017 IEEE Conference on Computer Vision and Pattern Recognition.Washington D.C., USA:IEEE Press, 2017:1522-1530.
[30] CHUN L Z, DIAN L, ZHI J Y, et al.YOLOv3:face detection in complex environments[J].International Journal of Computational Intelligence Systems, 2020, 13(1):1153.
[31] ZHANG K P, ZHANG Z P, LI Z F, et al.Joint face detection and alignment using multitask cascaded convolutional networks[J].IEEE Signal Processing Letters, 2016, 23(10):1499-1503.
[32] DENG J K, GUO J, VERVERAS E, et al.RetinaFace:single-shot multi-level face localisation in the wild[C]//Proceedings of 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition.Washington D.C., USA:IEEE Press, 2020:14-21.
[33] GUO J, DENG J K, LATTAS A, et al.Sample and computation redistribution for efficient face detection[EB/OL].[2021-10-05].https://arxiv.org/abs/2105.04714.

选择文件类型/文献管理软件名称

选择包含的内容