GFP-GAN: Towards Real-World Blind Face Restoration

with Generative Facial Prior

Xintao Wang      Yu Li      Honglun Zhang       Ying Shan
Applied Research Center (ARC), Tencent PCG

Comparisons with state-of-the-art face restoration methods: HiFaceGAN, DFDNet, Wan et al. and PULSE on the real-world low-quality images. While previous methods struggle to restore faithful facial details or retain face identity, our proposed GFP-GAN achieves a good balance of realness and fidelity with much less artifacts. In addition, the powerful generative facial prior allows us to perform restoration and color enhancement jointly.


Blind face restoration usually relies on facial priors, such as facial geometry prior or reference prior, to restore realistic and faithful details. However, very low-quality inputs cannot offer accurate geometric prior while high-quality references are inaccessible, limiting the applicability in real-world scenarios. In this work, we propose GFP-GAN that leverages rich and diverse priors encapsulated in a pretrained face GAN for blind face restoration. This Generative Facial Prior (GFP) is incorporated into the face restoration process via novel channel-split spatial feature transform layers, which allow our method to achieve a good balance of realness and fidelity. Thanks to the powerful generative facial prior and delicate designs, our GFP-GAN could jointly restore facial details and enhance colors with just a single forward pass, while GAN inversion methods require expensive image-specific optimization at inference. Extensive experiments show that our method achieves superior performance to prior art on both synthetic and real-world datasets.





          author = {Xintao Wang and Yu Li and Honglun Zhang and Ying Shan},
          title = {Towards Real-World Blind Face Restoration with Generative Facial Prior},
          booktitle={The IEEE Conference on Computer Vision and Pattern Recognition (CVPR)},
          year = {2021}


If you have any question, please contact Xintao Wang at