基于高分辨率网络的轻量型人体姿态估计研究

doi:10.19678/j.issn.1000-3428.0065262

摘要/Abstract

摘要： 现有人体姿态估计网络通常采用增加网络模型深度的方式来提高预测精度，但是导致网络模型的参数量以及运算复杂度增加。为此，在高分辨率网络的基础上提出一种融入Ghost模块、Sandglass模块以及注意力机制的轻量型人体姿态估计网络GSENet。参考基础残差模块Bottleneck以及Basicblock，将Bottleneck模块中的标准卷积替换为Ghost卷积，并且将Basicblock模块中的卷积替换为Sandglass模块，通过这种方式重新构建基础模块GSEneck以及GSEblock。在此基础上，加入注意力机制以保证网络的预测精度。实验结果表明，相比HRNet，GSENet在COCO数据集上的参数量和运算复杂度分别减少84.6%和76.1%，在MPII数据集上的参数量和运算复杂度降低84.6%和76.8%，在保持一定预测精度的情况下，GSENet网络模型能够有效地减少网络参数量并降低运算复杂度。

关键词: 人体姿态估计, 高分辨率网络, 轻量型网络, 注意力机制, 深度卷积神经网络

Abstract: The existing human pose estimation network improves the prediction accuracy by increasing the depth of the network model, which leads to an increase in the number of parameters and computational complexity of the model. Therefore, a lightweight human pose estimation network GSENet is proposed, based on a high-resolution network, integrating the Ghost module, Sandglass module, and attention mechanism.Referring to the basic residual modules Bottleneck and Basicblock, the standard convolution in Bottleneck is replaced by Ghost convolution, and the convolution in Basicblock is replaced by the Sandglass module.The basic modules GSEneck and GSEblock are rebuilt so as to reduce the number of parameters and the complexity of the calculation. An attention mechanism is added to ensure the prediction accuracy of the network.The experimental results show that compared with HRNet, the number of parameters and computational complexity of GSENet are reduced by 84.6% and 76.1%, respectively, on the COCO dataset and by 84.6% and 76.8%, respectively, on the MPII dataset.The GSENet network model can effectively reduce the number of network parameters and computational complexity while maintaining a certain prediction accuracy.

Key words: human pose estimation, high-resolution network, lightweight network, attention mechanism, deep Convolutional Neural Network(CNN)

中图分类号:

TP391.4

钟宝荣, 吴夏灵. 基于高分辨率网络的轻量型人体姿态估计研究[J]. 计算机工程, 2023, 49(4): 226-232,239.

ZHONG Baorong, WU Xialing. Research on Lightweight Human Pose Estimation Based on High-Resolution Network[J]. Computer Engineering, 2023, 49(4): 226-232,239.

http://www.ecice06.com/CN/Y2023/V49/I4/226

图/表 14

20230417190021

20230417190024

20230417190030

20230417190033

20230417190036

20230417190039

20230417190042

20230417190045

20230417190048

20230417190051

20230417190054

20230417190057

20230417190101

20230417190104

参考文献

[1] 段品生, 周建亮.基于姿态特征的建筑工人不安全行为刻画方法[J].安全与环境工程, 2022, 29(3):1-8. DUAN P S, ZHOU J L.Description method of construction workers' unsafe behaviors based on posture characteristics[J].Safety and Environmental Engineering, 2022, 29(3):1-8.(in Chinese)
[2] 马淼, 李贻斌, 武宪青, 等.视频中多特征融合人体姿态跟踪[J].中国图象图形学报, 2020, 25(7):1459-1472. MA M, LI Y B, WU X Q, et al.Human pose tracking based on multi-feature fusion in videos[J].Journal of Image and Graphics, 2020, 25(7):1459-1472.(in Chinese)
[3] 宋一凡, 张鹏, 刘立波.基于视觉手势识别的人机交互系统[J].计算机科学, 2019, 46(S2):570-574. SONG Y F, ZHANG P, LIU L B.Human-machine interaction system with vision-based gesture recognition[J].Computer Science, 2019, 46(S2):570-574.(in Chinese)
[4] 刘勇, 李杰, 张建林, 等.基于深度学习的二维人体姿态估计研究进展[J].计算机工程, 2021, 47(3):1-16. LIU Y, LI J, ZHANG J L, et al.Research progress of two-dimensional human pose estimation based on deep learning[J].Computer Engineering, 2021, 47(3):1-16.(in Chinese)
[5] FISCHLER M, ELSCHLAGER R.The representation and matching of pictorial structures[J].IEEE Transactions on Computers, 1973, 22(1):67-92.
[6] LECUN Y, BOTTOU L, BENGIO Y, et al.Gradient-based learning applied to document recognition[J].Proceedings of the IEEE, 1998, 86(11):2278-2324.
[7] TOSHEV A, SZEGEDY C.DeepPose:human pose estimation via deep neural networks[C]//Proceedings of Conference on Computer Vision and Pattern Recognition.Washington D.C., USA:IEEE Press, 2014:1653-1660.
[8] WEI S H, RAMAKRISHNA V, KANADE T, et al.Convolutional pose machines[C]//Proceedings of Conference on Computer Vision and Pattern Recognition.Washington D.C., USA:IEEE Press, 2016:4724-4732.
[9] NEWELL A, YANG K Y, DENG J.Stacked hourglass networks for human pose estimation[EB/OL].[2022-06-12].https://link.springer.com/content/pdf/10.1007/978-3-319-46484-8_29.pdf.
[10] SUN K, XIAO B, LIU D, et al.Deep high-resolution representation learning for human pose estimation[C]//Proceedings of Conference on Computer Vision and Pattern Recognition.Washington D.C., USA:IEEE Press, 2020:5686-5696.
[11] LIU S G, LI Y, HUA G G.Human pose estimation in video via structured space learning and halfway temporal evaluation[J].IEEE Transactions on Circuits and Systems for Video Technology, 2019, 29(7):2029-2038.
[12] CHENG B W, XIAO B, WANG J D, et al.HigherHRNet:scale-aware representation learning for bottom-up human pose estimation[C]//Proceedings of Conference on Computer Vision and Pattern Recognition.Washington D.C., USA:IEEE Press, 2020:5385-5394.
[13] GENG Z G, SUN K, XIAO B, et al.Bottom-up human pose estimation via disentangled keypoint regression[C]//Proceedings of Conference on Computer Vision and Pattern Recognition.Washington D.C., USA:IEEE Press, 2021:14671-14681.
[14] XIAO B, WU H P, WEI Y C.Simple baselines for human pose estimation and tracking[EB/OL].[2022-06-12].https://arxiv.org/pdf/1804.06208.pdf.
[15] ZHOU D Q, HOU Q B, CHEN Y P, et al.Rethinking bottleneck structure for efficient mobile network design[C]//Proceedings of European Conference on Computer Vision.Berlin, Germany:Springer, 2020:680-697.
[16] HAN K, WANG Y H, TIAN Q, et al.GhostNet:more features from cheap operations[C]//Proceedings of Conference on Computer Vision and Pattern Recognition.Washington D.C., USA:IEEE Press, 2020:1577-1586.
[17] YU C Q, XIAO B, GAO C X, et al.Lite-HRNet:a lightweight high-resolution network[C]//Proceedings of Conference on Computer Vision and Pattern Recognition.Washington D.C., USA:IEEE Press, 2021:10435-10445.
[18] WANG Q L, WU B G, ZHU P F, et al.ECA-net:efficient channel attention for deep convolutional neural networks[C]//Proceedings of Conference on Computer Vision and Pattern Recognition.Washington D.C., USA:IEEE Press, 2020:11531-11539.
[19] KRIZHEVSKY A, SUTSKEVER I, HINTON G E.ImageNet classification with deep convolutional neural networks[J].Communications of the ACM, 2017, 60(6):84-90.
[20] WANG P S, CHENG J.Accelerating convolutional neural networks for mobile applications[C]//Proceedings of the International Multimedia Conference.New York, USA:ACM Press, 2016:1-10.
[21] JADERBERG M, SIMONYAN K, ZISSERMAN A, et al.Spatial Transformer networks[EB/OL].[2022-6-12].https://arxiv.org/pdf/1506.02025.pdf.
[22] HU J, SHEN L, SUN G.Squeeze-and-excitation networks[C]//Proceedings of Conference on Computer Vision and Pattern Recognition.Washington D.C., USA:IEEE Press, 2018:7132-7141.
[23] HE K M, ZHANG X Y, REN S Q, et al.Deep residual learning for image recognition[C]//Proceedings of Conference on Computer Vision and Pattern Recognition.Washington D.C., USA:IEEE Press, 2016:770-778.
[24] SANDLER M, HOWARD A, ZHU M L, et al.MobileNetV2:inverted residuals and linear bottlenecks[C]//Proceedings of Conference on Computer Vision and Pattern Recognition.Washington D.C., USA:IEEE Press, 2018:4510-4520.
[25] LIN T Y, MAIRE M, BELONGIE S, et al.Microsoft COCO:common objects in context[EB/OL].[2022-06-12].https://link.springer.com/content/pdf/10.1007/978-3-319-10602-1_48.pdf.
[26] ANDRILUKA M, PISHCHULIN L, GEHLER P, et al.2D human pose estimation:new benchmark and state of the art analysis[C]//Proceedings of Conference on Computer Vision and Pattern Recognition.Washington D.C., USA:IEEE Press, 2014:3686-3693.

选择文件类型/文献管理软件名称

选择包含的内容