Home

Awesome

GAN Prior Embedded Network for Blind Face Restoration in the Wild

Paper | Supplementary | Demo | ModelScope

<a href="https://replicate.ai/yangxy/gpen"><img src="https://img.shields.io/static/v1?label=Replicate&message=Demo and Docker Image&color=blue"></a> Hugging Face Spaces

Tao Yang<sup>1</sup>, Peiran Ren<sup>1</sup>, Xuansong Xie<sup>1</sup>, Lei Zhang<sup>1,2</sup>
<sup>1</sup>DAMO Academy, Alibaba Group, Hangzhou, China
<sup>2</sup>Department of Computing, The Hong Kong Polytechnic University, Hong Kong, China

Face Restoration

<img src="figs/real_00.png" width="390px"/> <img src="figs/real_01.png" width="390px"/> <img src="figs/real_02.png" width="390px"/> <img src="figs/real_03.png" width="390px"/>

<img src="figs/Solvay_conference_1927_comp.jpg" width="784px"/>

Selfie Restoration

<img src="figs/selfie_00.jpg" width="390px"/> <img src="figs/selfie_01.jpg" width="390px"/>

Face Colorization

<img src="figs/colorization_00.jpg" width="390px"/> <img src="figs/colorization_01.jpg" width="390px"/>

Face Inpainting

<img src="figs/inpainting_00.jpg" width="390px"/> <img src="figs/inpainting_01.jpg" width="390px"/>

Conditional Image Synthesis (Seg2Face)

<img src="figs/seg2face_00.jpg" width="390px"/> <img src="figs/seg2face_01.jpg" width="390px"/>

News

(2023-02-15) GPEN-BFR-1024 and GPEN-BFR-2048 are now publicly available. Please download them via [ModelScope2].

(2023-02-15) We provide online demos via [ModelScope1] and [ModelScope2].

(2022-05-16) Add x1 sr model. Add --tile_size to avoid OOM.

(2022-03-15) Add x4 sr model. Try --sr_scale.

(2022-03-09) Add GPEN-BFR-2048 for selfies. <font color='red'>I have to take it down due to commercial issues. Sorry about that.</font>

(2021-12-29) Add online demos <a href="https://replicate.ai/yangxy/gpen"><img src="https://img.shields.io/static/v1?label=Replicate&message=Demo and Docker Image&color=blue"></a> Hugging Face Spaces. Many thanks to CJWBW and AK391.

(2021-12-16) Release a simplified training code of GPEN. It differs from our implementation in the paper, but could achieve comparable performance. We strongly recommend to change the degradation model.

(2021-12-09) Add face parsing to better paste restored faces back.

(2021-12-09) GPEN can run on CPU now by simply discarding --use_cuda.

(2021-12-01) GPEN can now work on a Windows machine without compiling cuda codes. Please check it out. Thanks to Animadversio. Alternatively, you can try GPEN-Windows. Many thanks to Cioscos.

(2021-10-22) GPEN can now work with SR methods. A SR model trained by myself is provided. Replace it with your own model if necessary.

(2021-10-11) The Colab demo for GPEN is available now <a href="https://colab.research.google.com/drive/1fPUsJCpQipp2Z5B5GbEXqpBGsMp-nvjm?usp=sharing"><img src="https://colab.research.google.com/assets/colab-badge.svg" alt="google colab logo"></a>.

Download models from Modelscope

pip install "modelscope[cv]" -f https://modelscope.oss-cn-beijing.aliyuncs.com/releases/repo.html

import cv2
from modelscope.pipelines import pipeline
from modelscope.utils.constant import Tasks
from modelscope.outputs import OutputKeys

portrait_enhancement = pipeline(Tasks.image_portrait_enhancement, model='damo/cv_gpen_image-portrait-enhancement-hires')
result = portrait_enhancement('https://modelscope.oss-cn-beijing.aliyuncs.com/test/images/marilyn_monroe_4.jpg')
cv2.imwrite('result.png', result[OutputKeys.OUTPUT_IMG])

It will automatically download the GPEN models. You can find the model in the local path ~/.cache/modelscope/hub/damo. Please note pytorch_model.pt, pytorch_model-2048.pt are respectively the 1024 and 2048 versions.

Usage

python pytorch cuda driver gcc

git clone https://github.com/yangxy/GPEN.git
cd GPEN

python demo.py --task FaceEnhancement --model GPEN-BFR-512 --in_size 512 --channel_multiplier 2 --narrow 1 --use_sr --sr_scale 4 --use_cuda --save_face --indir examples/imgs --outdir examples/outs-bfr

python demo.py --task FaceColorization --model GPEN-Colorization-1024 --in_size 1024 --use_cuda --indir examples/grays --outdir examples/outs-colorization

python demo.py --task FaceInpainting --model GPEN-Inpainting-1024 --in_size 1024 --use_cuda --indir examples/ffhq-10 --outdir examples/outs-inpainting

python demo.py --task Segmentation2Face --model GPEN-Seg2face-512 --in_size 512 --use_cuda --indir examples/segs --outdir examples/outs-seg2face

CUDA_VISIBLE_DEVICES='0,1,2,3' python -m torch.distributed.launch --nproc_per_node=4 --master_port=4321 train_simple.py --size 1024 --channel_multiplier 2 --narrow 1 --ckpt weights --sample results --batch 2 --path your_path_of_croped+aligned_hq_faces (e.g., FFHQ)

When testing your own model, set --key g_ema.

Please check out run.sh for more details.

Main idea

<img src="figs/architecture.png" width="784px"/>

Citation

If our work is useful for your research, please consider citing:

@inproceedings{Yang2021GPEN,
    title={GAN Prior Embedded Network for Blind Face Restoration in the Wild},
    author={Tao Yang, Peiran Ren, Xuansong Xie, and Lei Zhang},
    booktitle={IEEE Conference on Computer Vision and Pattern Recognition (CVPR)},
    year={2021}
}

License

© Alibaba, 2021. For academic and non-commercial use only.

Acknowledgments

We borrow some codes from Pytorch_Retinaface, stylegan2-pytorch, Real-ESRGAN, and GFPGAN.

Contact

If you have any questions or suggestions about this paper, feel free to reach me at yangtao9009@gmail.com.