Towards Real-World Blind Face Restoration with Generative Facial Prior

TL;DR

GFP-GAN introduces a novel approach to blind face restoration by leveraging a pretrained face GAN as a generative prior, enabling high-quality restoration from low-quality inputs without expensive optimization.

Contribution

The paper proposes GFP-GAN, a new method that incorporates a pretrained face GAN as a prior into face restoration, improving realness and fidelity in real-world scenarios.

Findings

Outperforms prior methods on synthetic datasets

Effective in real-world low-quality face restoration

Achieves high fidelity and color enhancement in a single pass

Abstract

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.