A multi-view SMPL fitting based on smplify-x
Windows or Linux, Python3
conda install pytorch torchvision torchaudio cudatoolkit=10.2 -c pytorch
pip install -r requirements.txt
Download the official SMPL model from SMPLify website (netural) and SMPL website (male/female). Then, rename the .pkl files and put them in the models/smpl folder. (see models/smpl/readme.txt)
Run python code/main.py --config cfg_files/fit_smpl.yaml
To apply our method to custom data, download the trained models from Baidu Netdisk or Google Drive and run AlphaPose to obtain 2D keypoints in JSON format. The keypoints are saved in data/keypoints
python code/keypoint_predict.py
We add a collision term based on SDF. You need to install sdf and set interpenetration: true in the cfg_files/fit_smpl.yaml before using this code.
cd sdf
python setup.py install
If the code is helpful in your research, please consider citing the following works.
@inproceedings{zhang2020object,
title={Object-Occluded Human Shape and Pose Estimation From a Single Color Image},
author={Zhang, Tianshu and Huang, Buzhen and Wang, Yangang},
booktitle={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition},
pages={7376--7385},
year={2020}
}
@inproceedings{SMPL-X:2019,
title = {Expressive Body Capture: 3D Hands, Face, and Body from a Single Image},
author = {Pavlakos, Georgios and Choutas, Vasileios and Ghorbani, Nima and Bolkart, Timo and Osman, Ahmed A. A. and Tzionas, Dimitrios and Black, Michael J.},
booktitle = {Proceedings IEEE Conf. on Computer Vision and Pattern Recognition (CVPR)},
year = {2019}
}