Double-chain Constraints for 3D Human Pose Estimation in Images and Videos,
Hongbo Kang, Yong Wang, Mengyuan Liu, Doudou Wu, Peng Liu, Wenming Yang
arXiv, 2023
This link contains the code that supports our latest work: DRPose
Protocol 1 (mean per-joint position error) when 2D keypoints detected by CPN, HRNet and the ground truth of 2D poses.
| Method | 2D Pose | MPJPE |
|---|---|---|
| DC-GCT | GT | 32.4 mm |
| DC-GCT | CPN | 48.4 mm |
| DC-GCT (w/refine) | CPN | 47.4 mm |
| DC-GCT | HRNet | 47.2 mm |
| DC-GCT (w/refine) | HRNet | 46.1 mm |
- Python 3.7+
- PyTorch >= 1.10.0
pip install -r requirement.txtPlease download the dataset here and refer to VideoPose3D to set up the Human3.6M dataset ('./dataset' directory).
${POSE_ROOT}/
|-- dataset
| |-- data_3d_h36m.npz
| |-- data_2d_h36m_gt.npz
| |-- data_2d_h36m_cpn_ft_h36m_dbb.npzThe pretrained model is here, please download it and put it in the './ckpt/pretrained' directory.
To test on Human3.6M on single frame, run:
python main.py --reload --previous_dir "ckpt/pretrained"
To train on Human3.6M with single frame, run:
python main.py --train -n 'name'
To begin, download the YOLOv3 and HRNet pretrained models here and put it in the './demo/lib/checkpoint' directory. Next, download the pretrained model and put it in the './ckpt/pretrained' directory. Lastly, Put your own images in the './demo/figure', and run:
python demo/vis.py
If you find our work useful in your research, please consider citing:
@article{kang2023double,
title={Double-chain Constraints for 3D Human Pose Estimation in Images and Videos},
author={Kang, Hongbo and Wang, Yong and Liu, Mengyuan and Wu, Doudou and Liu, Peng and Yang, Wenming},
journal={arXiv preprint arXiv:2308.05298},
year={2023}
}
Our code is extended from the following repositories. We thank the authors for releasing the codes.