Paper and Code

Inspired by sczhou Awesome-Face-Restoration and TaoWangzj Awesome-Face-Restoration.

Recommend

Several popular methods that perform well in my own test data.

Year	Pub	Abbreviation	Recommend
2021	CVPR	GFPGAN v1.4	⭐️⭐️⭐️⭐️⭐️
2022	NeurIPS	CodeFormer	⭐️⭐️⭐️⭐️⭐️
2021	CVPR	GPEN	⭐️⭐️⭐️

Blind Face Restoration (Image)

Geometric Prior

Year	Pub	Title	Links	Arch	Abbreviation
2018	CVPR	FSRNet: End-to-End Learning Face Super-Resolution with Facial Priors	[paper][code]	CNN	FSRNet
2021	TPAMI	Face Restoration via Plug-and-Play 3D Facial Priors	[paper][code]	CNN	-
2021	CVPR	Progressive Semantic-Aware Style Transformation for Blind Face Restoration	[paper][code]	GAN	PSFRGAN
2023	TNNLS	Multi-prior learning via neural architecture search for blind face restoration	[paper][code]	CNN	MFPSNet
2022	CVPR	Blind Face Restoration via Integrating Face Shape and Generative Priors	[paper][code]	GAN	SGPN

Reference Prior

Year	Pub	Title	Links	Arch	Abbreviation
2018	ECCV	Learning Warped Guidance for Blind Face Restoration	[paper][code]	CNN	GFRNet
2020	ECCV	Blind Face Restoration via Deep Multi-scale Component Dictionaries	[paper][code]	CNN	DFDNet
2020	CVPR	Enhanced Blind Face Restoration with Multi-Exemplar Images and Adaptive Spatial Feature Fusion	[paper][code]	CNN	ASFFNet
2022	TPAMI	Learning Dual Memory Dictionaries for Blind Face Restoration	[paper][code]	CNN	DMDNet
2022	ACM	MyStyle: A Personalized Generative Prior	[paper][code]	GAN	MyStyle
2023	arXiv	Personalized Restoration via Dual-Pivot Tuning	[paper][code]	Diffusion	Dual-Pivot Tuning
2024	arXiv	InstantRestore: Single-Step Personalized Face Restoration with Shared-Image Attention	[paper][code]	Diffusion	InstantRestore
2024	CVPR	PFStorer: Personalized Face Restoration and Super-Resolution	[paper][code]	Diffusion	PFStorer
2024	arXiv	RestorerID: Towards Tuning-Free Face Restoration with ID Preservation	[paper][code]	Diffusion	RestorerID
2024	arXiv	Overcoming False Illusions in Real-World Face Restoration with Multi-Modal GuidedDiffusion Model	[paper][code]	Diffusion	MGFR
2024	arXiv	Copy or Not? Reference-Based Face Image Restoration with Fine Details	[paper][code]	Diffusion	RefineFIR
2024	Sensors	PDGrad: Guiding Diffusion Model for Reference-Based Blind Face Restoration with Pivot Direction Gradient Guidance	[paper][code]	Diffusion	PDGrad
2024	CVPR	Restoration by Generation with Constrained Priors	[paper][code]	Diffusion	gen2res
2024	WACV	ENTED: Enhanced Neural Texture Extraction and Distribution for Reference-based Blind Face Restoration	[paper][code]	Diffusion	ENTED
2025	AAAI	FaceMe: Robust Blind Face Restoration with Personal Identification	[paper][code]	Diffusion	FaceMe

Generative Prior (good)

Year	Pub	Title	Links	Arch	Abbreviation
2020	MM	HiFaceGAN: Face Renovation via Collaborative Suppression and Replenishment	[paper][code]	GAN	HiFaceGAN
2020	CVPR	Image Processing Using Multi-Code GAN Prior	[paper][code]	GAN	mGANprior
2021	CVPR	Towards Real-World Blind Face Restoration with Generative Facial Prior	[paper][code]	GAN	GFPGAN
2021	CVPR	GAN Prior Embedded Network for Blind Face Restoration in the Wild	[paper][code]	GAN	GPEN
2022	CVPR	Blind Face Restoration via Integrating Face Shape and Generative Priors	[paper][code]	GAN	SGPN
2022	CVPR	GCFSR: a Generative and Controllable Face Super Resolution Method Without Facial and GAN Priors	[paper][code]	GAN	GCFSR
2022	AAAI	Panini-Net: GAN Prior Based Degradation-Aware Feature Interpolation for Face Restoration	[paper][code]	GAN	Panini-Net
2022	Arxiv	FaceFormer: Scale-aware Blind Face Restoration with Transformers	[paper][code]	ViT	FaceFormer
2022	Arxiv	Blind Face Restoration: Benchmark Datasets and a Baseline Model	[paper][code]	ViT	STUNet
2023	CVPR	TFRGAN: Leveraging Text Information for Blind Face Restoration with Extreme Degradation	[paper][code]	GAN	TFRGAN
2023	TCSVT	DEAR-GAN: Degradation-Aware Face Restoration With GAN Prior	[paper][code]	GAN	DEAR-GAN
2023	ACMMM	Exploring Correlations in Degraded Spatial Identity Features for Blind Face Restoration	[paper][code]	GAN	MemGAN
2023	FSP	Degradation Learning and Skip-Transformer for Blind Face Restoration	[paper][code]	GAN	-
2023	ACMMM	Exploring Correlations in Degraded Spatial Identity Features for Blind Face Restoration	[paper][code]	GAN	-
2024	AAAI	Blind Face Restoration under Extreme Conditions: Leveraging 3D-2D Prior Fusion for Superior Structural and Texture Recovery	[paper][code]	GAN	FREx

Vector Quantized Codebook Prior (good)

Year	Pub	Title	Links	Arch	Abbreviation
2022	CVPR	RestoreFormer: High-Quality Blind Face Restoration from Undegraded Key-Value Pairs	[paper][code]	ViT	RestoreFormer
2022	NeurIPS	Towards Robust Blind Face Restoration with Codebook Lookup Transformer	[paper][code]	ViT	CodeFormer
2022	ECCV	VQFR: Blind Face Restoration with Vector-Quantized Dictionary and Parallel Decoder	[paper][code]	CNN	VQFR
2023	TPAMI	RestoreFormer++: Towards Real-World Blind Face Restoration from Undegraded Key-Value Pairs	[paper][code]	ViT	RestoreFormer++
2022	CVPR	Rethinking Deep Face Restoration	[paper][code]	GAN	-
2024	ICLR	Dual Associated Encoder for Face Restoration	[paper][code]	ViT	DAEFR

Diffusion Prior

Year	Pub	Title	Links	Arch	Abbreviation
2022	Arxiv	Difface: Blind Face Restoration with Diffused Error Contraction	[paper][code]	Diffusion	DifFace
2023	CVPR	DR2: Diffusion-based Robust Degradation Remover for Blind Face Restoration	[paper][code]	Diffusion	DR2
2023	WACV	AT-DDPM: Restoring Faces degraded by Atmospheric Turbulence using Denoising Diffusion Probabilistic Models	[paper][code]	Diffusion	AT-DDPM
2023	ACMMM	DiffBFR: Bootstrapping Diffusion Model for Blind Face Restoration	[paper][code]	Diffusion	DiffBFR
2023	NeurIPS	PGDiff: Guiding Diffusion Models for Versatile Face Restoration via Partial Guidance	[paper][code]	Diffusion	PGDiff
2024	TCSVT	Towards Real-World Blind Face Restoration with Generative Diffusion Prior	[paper][code]	Diffusion	BFRffusion
2024	WACV	Diffuse and Restore: A Region-Adaptive Diffusion Model for Identity-Preserving Blind Face Restoration	[paper][code]	Diffusion	-
2024	ECCV	DiffBIR: Towards Blind Image Restoration with Generative Diffusion Prior	[paper][code]	Diffusion	DiffBIR
2024	NeurIPS	One-Step Effective Diffusion Network for Real-World Image Super-Resolution	[paper][code]	Diffusion	OSEDiff
2024	Arxiv	OSDFace: One-Step Diffusion Model for Face Restoration	[paper][code]	Diffusion	OSDFace

Others

Year	Pub	Title	Links	Arch	Abbreviation
2023	ICCV	Towards Authentic Face Restoration with Iterative Diffusion Models and Beyond	[paper][code]	Diffusion	IDM
2024	CVPR	WaveFace: Authentic Face Restoration with Efficient Frequency Recovery	[paper][code]	Diffusion	WaveFace
2024	Arxiv	CLR-Face: Conditional Latent Refinement for Blind Face Restoration Using Score-Based Diffusion Models	[paper][code]	Diffusion	CLR-Face
2024	WACV	Diffuse and Restore: A Region-Adaptive Diffusion Model for Identity-Preserving Blind Face Restoration	[paper][code]	Diffusion	Diffuse and Restore
2024	arXiv	AuthFace: Towards Authentic Blind Face Restoration with Face-oriented Generative Diffusion Prior	[paper][code]	Diffusion	AuthFace
2025	Arxiv	InfoBFR: Real-World Blind Face Restoration via Information Bottleneck	[paper][code]	Diffusion	InfoBFR
2025	WACV	Towards Unsupervised Blind Face Restoration using Diffusion Prior	[paper][code]	Diffusion	DT-BFR

Blind Face Restoration (Video)

Year	Pub	Title	Links	Arch	Abbreviation
2023	Arxiv	FLAIR: A Conditional Diffusion Framework with Applications to Face Video Restoration	[paper][code]	Diffusion	FLAIR
2024	IJCAI	Beyond Alignment: Blind Video Face Restoration via Parsing-Guided Temporal-Coherent Transformer	[paper][code]	ViT	PGTFormer
2024	ECCV	KEEP: Kalman-Inspired Feature Propagation for Video Face Super-Resolution	[paper][code]	ViT	KEEP
2024	TIP	FIR2FVR: Analysis and Benchmarking of Extending Blind Face Image Restoration to Video	[paper][code]	General	FIR2FVR
2024	ACML	Universal Video Face Restoration Method Based on Vision-Language Model	[paper][code]	Diffusion	-
2024	ACM MM	Blind Face Video Restoration with Temporal Consistent Generative Prior and Degradation-Aware Prompt	[paper][code]	Diffusion	StableBFVR
2025	CVPR	Efficient Video Face Enhancement with Enhanced Spatial-Temporal Consistency	[paper][code]	ViT	BFVR-STC
2025	ArXiv	KEEP: Kalman-Inspired Feature Propagation for Video Face Super-Resolution	[paper][code]	Diffusion	SVFR

Datasets

Image 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
EFHQ	512 x 512	Multi-purpose ExtremePose-Face-HQ dataset

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