mlsanityData to 7 Oct 2026

How far have we gone in Generative Image Restoration? A study on its capability, limitations and evaluation practices

arXiv 2603.05010 ↗First posted 5 Mar 2026cs.CVcited by 2

Not on the map. Why?

Abstract (arXiv)

Generative Image Restoration (GIR) has achieved impressive perceptual realism, but how far have its practical capabilities truly advanced compared with previous methods? To answer this, we present a large-scale study grounded in a new multi-dimensional evaluation pipeline that assesses models on detail, sharpness, semantic correctness, and overall quality. Our analysis covers diverse architectures, including diffusion-based, GAN-based, PSNR-oriented, and general-purpose generation models, revealing critical performance disparities. Furthermore, our analysis uncovers a key evolution in failure modes that signifies a paradigm shift for the perception-oriented low-level vision field. The central challenge is evolving from the previous problem of detail scarcity (under-generation) to the new frontier of detail quality and semantic control (preventing over-generation). We also leverage our benchmark to train a new IQA model that better aligns with human perceptual judgments. Ultimately, this work provides a systematic study of modern generative image restoration models, offering crucial insights that redefine our understanding of their true state and chart a course for future development.

Key facts. Cited by 2 papers in the map (data through 7 Oct 2026). Builds on 80 of its 108 references that are in the map. First posted on arXiv 5 Mar 2026.

Builds on

80 of 108 references.

References · 80

  1. 2018The Unreasonable Effectiveness of Deep Features as a Perceptual Metriccited by 11,029
  2. 2016Photo-Realistic Single Image Super-Resolution Using a Generative Adversarial Networkcited by 1,898
  3. 2021SwinIR: Image Restoration Using Swin Transformercited by 1,869
  4. 2017Enhanced Deep Residual Networks for Single Image Super-Resolutioncited by 1,543
  5. 2021Image Super-Resolution via Iterative Refinementcited by 1,506
  6. 2016Beyond a Gaussian Denoiser: Residual Learning of Deep CNN for Image Denoisingcited by 1,352
  7. 2018ESRGAN: Enhanced Super-Resolution Generative Adversarial Networkscited by 1,250
  8. 2021MUSIQ: Multi-scale Image Quality Transformercited by 1,058
  9. 2022Exploring CLIP for Assessing the Look and Feel of Imagescited by 974
  10. 2021Real-ESRGAN: Training Real-World Blind Super-Resolution with Pure Synthetic Datacited by 961
  11. 2025FLUX.1 Kontext: Flow Matching for In-Context Image Generation and Editing in Latent Spacecited by 960
  12. 2020Image Quality Assessment: Unifying Structure and Texture Similaritycited by 792
38 more
  1. 2016Deep Multi-scale Convolutional Neural Network for Dynamic Scene Deblurringcited by 722
  2. 2017The Perception-Distortion Tradeoffcited by 582
  3. 2018Real-world Anomaly Detection in Surveillance Videoscited by 570
  4. 2021Designing a Practical Degradation Model for Deep Blind Image Super-Resolutioncited by 524
  5. 2022MANIQA: Multi-dimension Attention Network for No-Reference Image Quality Assessmentcited by 514
  6. 2023Q-Align: Teaching LMMs for Visual Scoring via Discrete Text-Defined Levelscited by 489
  7. 2023Exploiting Diffusion Prior for Real-World Image Super-Resolutioncited by 486
  8. 2022Activating More Pixels in Image Super-Resolution Transformercited by 433
  9. 2017DeblurGAN: Blind Motion Deblurring Using Conditional Adversarial Networkscited by 366
  10. 2023ResShift: Efficient Diffusion Model for Image Super-resolution by Residual Shiftingcited by 360
  11. 2019KonIQ-10k: An ecologically valid database for deep learning of blind image quality assessmentcited by 351
  12. 2023SeeSR: Towards Semantics-Aware Real-World Image Super-Resolutioncited by 311
  13. 2024Scaling Up to Excellence: Practicing Model Scaling for Photo-Realistic Image Restoration In the Wildcited by 281
  14. 2023Blind Image Quality Assessment via Vision-Language Correspondence: A Multitask Learning Perspectivecited by 276
  15. 2023Pixel-Aware Stable Diffusion for Realistic Image Super-resolution and Personalized Stylizationcited by 251
  16. 2024One-Step Effective Diffusion Network for Real-World Image Super-Resolutioncited by 243
  17. 2016Deep Neural Networks for No-Reference and Full-Reference Image Quality Assessmentcited by 224
  18. 2023TextDiffuser: Diffusion Models as Text Painterscited by 223
  19. 2019From Patches to Pictures (PaQ-2-PiQ): Mapping the Perceptual Space of Picture Qualitycited by 221
  20. 2015Large-scale Classification of Fine-Art Paintings: Learning The Right Metric on The Right Featurecited by 181
  21. 2020PIPAL: a Large-Scale Image Quality Assessment Dataset for Perceptual Image Restorationcited by 177
  22. 2018PieAPP: Perceptual Image-Error Assessment through Pairwise Preferencecited by 158
  23. 2017RankIQA: Learning from Rankings for No-reference Image Quality Assessmentcited by 150
  24. 2020MetaIQA: Deep Meta-learning for No-Reference Image Quality Assessmentcited by 144
  25. 2016Learning a No-Reference Quality Metric for Single-Image Super-Resolutioncited by 143
  26. 2023AGIQA-3K: An Open Database for AI-Generated Image Quality Assessmentcited by 142
  27. 2025Teaching Large Language Models to Regress Accurate Image Quality Scores using Score Distributioncited by 139
  28. 2020Defocus Deblurring Using Dual-Pixel Datacited by 138
  29. 2016Clearing the Skies: A deep network architecture for single-image rain removalcited by 136
  30. 2020Uncertainty-Aware Blind Image Quality Assessment in the Laboratory and Wildcited by 134
  31. 2023Diffusion Models for Image Restoration and Enhancement: A Comprehensive Surveycited by 108
  32. 2023Depicting Beyond Scores: Advancing Image Quality Assessment through Multi-modal Language Modelscited by 104
  33. 2024Pixel-level and Semantic-level Adjustable Super-resolution: A Dual-LoRA Approachcited by 93
  34. 2025Q-Insight: Understanding Image Quality via Visual Reinforcement Learningcited by 93
  35. 2023Hierarchical Integration Diffusion Model for Realistic Image Deblurringcited by 87
  36. 2023Refusion: Enabling Large-Size Realistic Image Restoration with Latent-Space Diffusion Modelscited by 86
  37. 2021Iterative Filter Adaptive Network for Single Image Defocus Deblurringcited by 85
  38. 2024Arbitrary-steps Image Super-resolution via Diffusion Inversioncited by 85

50 most cited of 80 All 80 (JSON)

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2026: 2.

2 papers.

  1. 2026Can Nano Banana 2 Replace Traditional Image Restoration Models? An Evaluation of Its Performance on Image Restoration Taskscited by 4
  2. 2026LL-Bench: Rethinking Low-Level Vision Evaluation in the Era of Large-Scale Generative Modelscited by 0