3D-Aware Face Swapping
Official PyTorch implementation of the CVPR 2023 paper "3D-Aware Face Swapping"
3D-Aware Face Swapping
Yixuan Li, Chao Ma, Yichao Yan, Wenhan Zhu, Xiaokang Yang
Abstract: Face swapping is an important research topic in computer vision with wide applications in entertainment and privacy protection. Existing methods directly learn to swap 2D facial images, taking no account of the geometric information of human faces. In the presence of large pose variance between the source and the target faces, there always exist undesirable artifacts on the swapped face. In this paper, we present a novel 3D-aware face swapping method that generates high-fidelity and multi-view-consistent swapped faces from single-view source and target images. To achieve this, we take advantage of the strong geometry and texture prior of 3D human faces, where the 2D faces are projected into the latent space of a 3D generative model. By disentangling the identity and attribute features in the latent space, we succeed in swapping faces in a 3D-aware manner, being robust to pose variations while transferring fine-grained facial details. Extensive experiments demonstrate the superiority of our 3D-aware face swapping framework in terms of visual quality, identity similarity, and multi-view consistency. Project page: https://lyx0208.github.io/3dSwap
- Create and activate the Python environment:
conda create -n 3dSwap python=3.8conda activate 3dSwappip install -r requirements.txt
-
We preprocess the images from the original FFHQ and CelebA-HD dataset with the data preprocessing code from EG3D, including re-cropping the images and extracting according camera poses.
Download our pretrained model from Baidu Disk or Goole Drive. Put model_ir_se50.pth under the "models" folder and other files under the "checkpoints" folder.
Then run:
python run_3dSwap.pyIf you only want to perform the 3D GAN inversion without face swapping, run:
python run_inversion.pyFirst, download the preprocessed FFHQ dataset from here and put it under the "datasets" folder.
To train the inversion module, run:
python -m torch.distributed.launch --nproc_per_node=4 --master_port=12345 train_inversion.py --exp_dir=inversionTo train the faceswapping module, run:
python -m torch.distributed.launch --nproc_per_node=4 --master_port=12345 train_faceswap.py --exp_dir=faceswap@InProceedings{Li_2023_CVPR,
author = {Li, Yixuan and Ma, Chao and Yan, Yichao and Zhu, Wenhan and Yang, Xiaokang},
title = {3D-Aware Face Swapping},
booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
month = {June},
year = {2023},
pages = {12705-12714}
}
- Our code is developed based on:
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent