sczhou / awesome-face-restoration

A curated list of face restoration & enhancement papers and resources

Awesome Face Restoration & Enhancement

A curated list of awesome face restoration & enhancement papers and resources 🐳, inspired by awesome-NeRF.

Welcome to add papers and other resources related to this topic [submit a pull request] 🤗

Table of Contents

• Papers
  • Face Image Restoration
  • Face Video Restoration
• Datasets
  • High-Res Face Dataset
  • Low-Res Face Dataset
  • Video Face Dataset
  • Other Face Dataset

Papers

Face Image Restoration

Diffusion Model

• [CVPR 2023] DR2: Diffusion-based Robust Degradation Remover for Blind Face Restoration, Wang et al. Paper | Bibtex
• [Arxiv 2022] DifFace: Blind Face Restoration with Diffused Error Contraction, Yue et al. Paper | Github | Demo | Bibtex

Generative Prior - VQGAN

• [NeurIPS 2022] CodeFormer: Towards Robust Blind Face Restoration with Codebook Lookup Transformer, Zhou et al. Paper | Project | Github | Demo | Bibtex

• [ECCV 2022] VQFR: Blind Face Restoration with Vector-Quantized Dictionary and Parallel Decoder, Gu et al. Paper | Project | Github | Bibtex

• [CVPR 2022] RestoreFormer: High-Quality Blind Face Restoration from Undegraded Key-Value Pairs, Wang et al. Paper | Github | Bibtex

• [CVPR 2022] Rethinking Deep Face Restoration. Zhao et al. Paper | Bibtex

Generative Prior - StyleGAN2

• [CVPR 2021] GFPGAN: Towards Real-World Blind Face Restoration with Generative Facial Prior, Wang et al. Paper | Project | Github | Demo | Bibtex

• [CVPR 2021] GLEAN: Generative Latent Bank for Large-Factor Image Super-Resolution, Chan et al. Paper | Project | Github | Bibtex

• [CVPR 2021] GPEN: GAN Prior Embedded Network for Blind Face Restoration in the Wild, Yang et al. Paper | Github | Bibtex

GAN Inversion

• [CVPR 2020] PULSE: Self-Supervised Photo Upsampling via Latent Space Exploration of Generative Models, Menon et al. Paper | Github | Bibtex

Dictionary Learning

• [TPAMI 2022] DMDNet: Learning Dual Memory Dictionaries for Blind Face Restoration, Li et al. Paper | Github | CelebRef-HQ Dataset | Bibtex
• [ECCV 2020] DFDNet: Blind Face Restoration via Deep Multi-scale Component Dictionaries, Li et al. Paper | Github | Bibtex

Reference/Exemplar Prior

• [CVPR 2020] ASFFNet: Enhanced Blind Face Restoration With Multi-Exemplar Images and Adaptive Spatial Feature Fusion, Li et al. Paper | Github | Bibtex

Geometry Facial Prior

• [CVPR 2021] PSFRGAN: Progressive Semantic-Aware Style Transformation for Blind Face Restoration, Chen et al. Paper | Github | Bibtex

3D Face Shape Prior

• [CVPR 2022] SGPN: Blind Face Restoration via Integrating Face Shape and Generative Priors, Zhu et al. Paper | Bibtex

Personalized Restoration

• [ArXiv 2022] MyStyle: A Personalized Generative Prior, Nitzan et al. Paper | Github | Project | Bibtex

Others

• [TIP 2020] Learning Spatial Attention for Face Super-Resolution, Chen et al. Paper | Github | Bibtex

 

Face Video Restoration

• [CVPRW 2022] VFHQ: A High-Quality Dataset and Benchmark for Video Face Super-Resolution, Xie et al. Paper | Project | Bibtex

Datasets

High-Resolution Face Dataset

Dataset	Resolution	Description
FFHQ	1024 x 1024	7,0000 high-quality face images (usually used for training)
CelebA-HQ	1024 x 1024	3,0000 high-quality face images (usually used for evaluation)
CelebAMask-HQ	512 x 512	3,0000 face images with 19 facial classes
CelebRef-HQ	512 x 512	high-quality face images with multiple same-identity references

Low-Resolution Face Dataset

Dataset	Description
CelebA	a large-scale face attributes dataset with more than 200K celebrity images
WIDER-Test	970 real-world severely degraded face images from the WIDER Face dataset (for test)
LFW-Test	1711 real-world degraded faces collected from the LFW dataset (for test)
WebPhoto-Test	407 real-world degraded faces collected from the Internet (for test)
CelebChild-Test	180 real-world degraded child faces collected from the Internet (for test)

Video Face Dataset

Dataset	Description
TalkingHead-1KH	500k video clips, of which about 80k are greater than 512x512 resolution
VFHQ	16,000 high-fidelity clips of diverse interview scenarios
CelebV-HQ	35,666 video clips involving 15,653 identities and 83 manually labeled facial attributes
CelebV-Text	70,000 in-the-wild face video clips covering diverse visual content

Other Face Dataset

Dataset	Description
CelebA-Dialog	a large-scale visual-language face dataset with fine-grained labels and captions
CelebA-Spoof	a large-scale face anti-spoofing dataset with rich attributes and spoof types
PPR10K	a large-scale portrait photo retouching dataset

awesome-face-restoration's Issues

Add supir

https://supir.xpixel.group/ a new parer from the xpixel (diffbir is also from this group)

哪个速度快？