Inspired by sczhou Awesome-Face-Restoration and TaoWangzj Awesome-Face-Restoration.
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 | ⭐️⭐️⭐️ |
| 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] | GAN | SGPN |
| 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] | 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] | 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] | 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] | 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] | Diffusion | ENTED |
| 2025 | AAAI | FaceMe: Robust Blind Face Restoration with Personal Identification | [paper][code] | Diffusion | FaceMe |
| 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 |
| 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 |
| 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 |
| Year | Pub | Title | Links | Arch | Abbreviation |
|---|---|---|---|---|---|
| 2023 | ICCV | Towards Authentic Face Restoration with Iterative Diffusion Models and Beyond | [paper] | Diffusion | IDM |
| 2024 | CVPR | WaveFace: Authentic Face Restoration with Efficient Frequency Recovery | [paper] | Diffusion | WaveFace |
| 2024 | Arxiv | CLR-Face: Conditional Latent Refinement for Blind Face Restoration Using Score-Based Diffusion Models | [paper] | Diffusion | CLR-Face |
| 2024 | WACV | Diffuse and Restore: A Region-Adaptive Diffusion Model for Identity-Preserving Blind Face Restoration | [paper] | 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] | Diffusion | InfoBFR |
| 2025 | WACV | Towards Unsupervised Blind Face Restoration using Diffusion Prior | [paper][code] | Diffusion | DT-BFR |
| 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] | General | FIR2FVR |
| 2024 | ACML | Universal Video Face Restoration Method Based on Vision-Language Model | [paper] | Diffusion | - |
| 2024 | ACM MM | Blind Face Video Restoration with Temporal Consistent Generative Prior and Degradation-Aware Prompt | [paper] | 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 |
| 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 |
| 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 |
1 commits
Inspired by sczhou Awesome-Face-Restoration and TaoWangzj Awesome-Face-Restoration.
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 | ⭐️⭐️⭐️ |
| 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] | GAN | SGPN |
| 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] | 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] | 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] | 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] | 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] | Diffusion | ENTED |
| 2025 | AAAI | FaceMe: Robust Blind Face Restoration with Personal Identification | [paper][code] | Diffusion | FaceMe |
| 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 |
| 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 |
| 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 |
| Year | Pub | Title | Links | Arch | Abbreviation |
|---|---|---|---|---|---|
| 2023 | ICCV | Towards Authentic Face Restoration with Iterative Diffusion Models and Beyond | [paper] | Diffusion | IDM |
| 2024 | CVPR | WaveFace: Authentic Face Restoration with Efficient Frequency Recovery | [paper] | Diffusion | WaveFace |
| 2024 | Arxiv | CLR-Face: Conditional Latent Refinement for Blind Face Restoration Using Score-Based Diffusion Models | [paper] | Diffusion | CLR-Face |
| 2024 | WACV | Diffuse and Restore: A Region-Adaptive Diffusion Model for Identity-Preserving Blind Face Restoration | [paper] | 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] | Diffusion | InfoBFR |
| 2025 | WACV | Towards Unsupervised Blind Face Restoration using Diffusion Prior | [paper][code] | Diffusion | DT-BFR |
| 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] | General | FIR2FVR |
| 2024 | ACML | Universal Video Face Restoration Method Based on Vision-Language Model | [paper] | Diffusion | - |
| 2024 | ACM MM | Blind Face Video Restoration with Temporal Consistent Generative Prior and Degradation-Aware Prompt | [paper] | 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 |
| 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 |
| 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 |
1 commits