An Articulated Structure-aware Network for 3D Human Pose Estimation

Zhenhua Tang, Xiaoyan Zhang, Junhui Hou
Proceedings of The Eleventh Asian Conference on Machine Learning, PMLR 101:48-63, 2019.

Abstract

In this paper, we propose a new end-to-end articulated structure-aware network to regress 3D joint coordinates from the given 2D joint detections. The proposed method is capable of dealing with hard joints well that usually fail existing methods. Specifically, our framework cascades a refinement network with a basic network for two types of joints, and employs a attention module to simulate a camera projection model. In addition, we propose to use a random enhancement module to intensify the constraints between joints. Experimental results on the Human3.6M and HumanEva databases demonstrate the effectiveness and flexibility of the proposed network, and errors of hard joints and bone lengths are significantly reduced, compared with state-of-the-art approaches.

Cite this Paper


BibTeX
@InProceedings{pmlr-v101-tang19a, title = {An Articulated Structure-aware Network for 3D Human Pose Estimation}, author = {Tang, Zhenhua and Zhang, Xiaoyan and Hou, Junhui}, booktitle = {Proceedings of The Eleventh Asian Conference on Machine Learning}, pages = {48--63}, year = {2019}, editor = {Lee, Wee Sun and Suzuki, Taiji}, volume = {101}, series = {Proceedings of Machine Learning Research}, month = {17--19 Nov}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v101/tang19a/tang19a.pdf}, url = {https://proceedings.mlr.press/v101/tang19a.html}, abstract = {In this paper, we propose a new end-to-end articulated structure-aware network to regress 3D joint coordinates from the given 2D joint detections. The proposed method is capable of dealing with hard joints well that usually fail existing methods. Specifically, our framework cascades a refinement network with a basic network for two types of joints, and employs a attention module to simulate a camera projection model. In addition, we propose to use a random enhancement module to intensify the constraints between joints. Experimental results on the Human3.6M and HumanEva databases demonstrate the effectiveness and flexibility of the proposed network, and errors of hard joints and bone lengths are significantly reduced, compared with state-of-the-art approaches.} }
Endnote
%0 Conference Paper %T An Articulated Structure-aware Network for 3D Human Pose Estimation %A Zhenhua Tang %A Xiaoyan Zhang %A Junhui Hou %B Proceedings of The Eleventh Asian Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2019 %E Wee Sun Lee %E Taiji Suzuki %F pmlr-v101-tang19a %I PMLR %P 48--63 %U https://proceedings.mlr.press/v101/tang19a.html %V 101 %X In this paper, we propose a new end-to-end articulated structure-aware network to regress 3D joint coordinates from the given 2D joint detections. The proposed method is capable of dealing with hard joints well that usually fail existing methods. Specifically, our framework cascades a refinement network with a basic network for two types of joints, and employs a attention module to simulate a camera projection model. In addition, we propose to use a random enhancement module to intensify the constraints between joints. Experimental results on the Human3.6M and HumanEva databases demonstrate the effectiveness and flexibility of the proposed network, and errors of hard joints and bone lengths are significantly reduced, compared with state-of-the-art approaches.
APA
Tang, Z., Zhang, X. & Hou, J.. (2019). An Articulated Structure-aware Network for 3D Human Pose Estimation. Proceedings of The Eleventh Asian Conference on Machine Learning, in Proceedings of Machine Learning Research 101:48-63 Available from https://proceedings.mlr.press/v101/tang19a.html.

Related Material