JchenXu/motion-prior-reconstruction

Exploring Versatile Prior for Human Motion via Motion Frequency Guidance

This is the codebase for video-based human motion reconstruction in human-motion-prior.

Installation

Requirements

Python 3.6
PyTorch 1.1.0

Because this project is based on our pretrained human motion prior, please clone the prior repository and this repository as follows:

git clone https://github.com/JchenXu/human-motion-prior.git human_motion_prior                                                                                          
git clone https://github.com/JchenXu/motion-prior-reconstruction.git

and run the following command to install the dependencies:

pip install -r requirements.txt

Data Preparation

Please download the required data (i.e., our pre-trained prior model and SMPL model parameters) here, and then, uncompress and put it in data/mp_data.

Then, refer to this for data generation, and put all data files in data/mp_db.

The whole data directory is like:

motion-prior-reconstruction/data
├── mp_data
│   ├── ...
|   └── ...
|
├── mp_db
    ├── 3dpw_train_db.pt
    └── insta_train_db.h5
    └── ...

Training

Run the commands below to start training:

export PYTHONPATH=../human_motion_prior
python train.py --cfg configs/config_3dpw.yaml

Evaluation

Modify the trained checkpoint in configs to evaluation your trained model. Then, run the commands below to start evaluation:

export PYTHONPATH=../human_motion_prior
python eval.py --cfg configs/config_3dpw.yaml

Also, we provide our pre-trained checkpoint here.

Citation

@inproceedings{human_motion_prior,
  title = {Exploring Versatile Prior for Human Motion via Motion Frequency Guidance},
  author = {Jiachen Xu, Min Wang, Jingyu Gong, Wentao Liu, Chen Qian, Yuan Xie, Lizhuang Ma},
  booktitle = {2021 international conference on 3D vision (3DV)},
  year = {2021}
}

Acknowledgement

We thank the authors of VIBE for their released code, and this base codes are largely borrowed from them.

JchenXu / motion-prior-reconstruction