DiffFace: Diffusion-based Face Swapping with Facial Guidance

Official PyTorch implementation of Diffusion-based Face Swapping with Facial Guidance

Environment setup

git clone https://github.com/hxngiee/DiffFace.git
cd DiffFace
conda create -n DiffFace python=3.9.7
conda activate DiffFace

conda install pytorch==1.10.0 torchvision==0.11.0 torchaudio==0.10.0 -c pytorch 

pip install -r requirements.txt

Download Pretrained Weights

The weights required for the execution of our DiffFace can be downloaded from link.

mkdir checkpoints
mv Arcface.tar checkpoints/ 
mv FaceParser.pth checkpoints/ 
mv GazeEstimator.pt checkpoints/
mv Model.pt checkpoints/

Directories structure

The dataset and checkpoints should be placed in the following structures below

DiffFace
├── checkpoints
    ├── Arcface.tar
    ├── FaceParser.pth
    ├── GazeEstimator.pt
    └── Model.pt
├── data
    └── src
        ├── 001.png
        └── ...
    └── targ
        ├── 001.png
        └── ...
├── models
├── optimization
├── utils
└── main.py

Quick Inference Using Pretrained Model

Place source and target images in data/src, and data/targ. Then run the following.

python main.py --output_path output/example

License

This licence allows for academic and non-commercial purpose only. The entire project is under the CC-BY-NC 4.0 license.

Citation

If you find DiffFace useful for your work please cite:

@inproceedings{kim2022diffface,
  title = {DiffFace: Diffusion-based Face Swapping with Facial Guidance},
  author = {K. Kim, Y. Kim, S. Cho, J. Seo, J. Nam, K. Lee, S. Kim, K. Lee},
  journal = {Arxiv},
  year = {2022}
}

Acknowledgments

This code borrows heavily from Blended-Diffusion and Guided-Diffusion.

syunar/DiffFace