Research map / Image generation / Image restoration
Restore, Assess, Repeat: A Unified Framework for Iterative Image Restoration
I-Hsiang Chen, Isma Hadji, Enrique Sanchez, Adrian Bulat, Sy-Yen Kuo, Radu Timofte, Georgios Tzimiropoulos, Brais Martinez
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| replace CLIP IQA with DepictQA free-form IQA (ablation) | SD1.5 restorer · UDC unknown | PSNR: 20.67 -> 22.34 | Improved |
| upgrade backbone from SD1.5 to SD3.5 (ablation) | RAR · UDC unknown | PSNR: 22.34 -> 24.41 | Improved |
| upgrade backbone from SD1.5 to SD3.5 (ablation) | RAR · Composite Group C | PSNR: 17.67 -> 17.88 | No clear effect |
| add noise-free direct degraded-to-clean mapping (ablation) | RAR SD3.5 · UDC unknown | PSNR: 24.41 -> 24.70 | Improved |
| feed IQA in latent space instead of pixel space (ablation) | RAR SD3.5 · Composite Group C | PSNR: 17.89 -> 18.72 | Improved |
| condition on IQA embeddings instead of decoded text (ablation) | RAR SD3.5 · UDC unknown | PSNR: 24.90 -> 28.49 | Improved |
| condition on IQA embeddings instead of decoded text (ablation) | RAR SD3.5 · Composite Group C | LPIPS: 0.1576 -> 0.1730 | Hurt |
| add iterative conditioning during training (ablation) | SD1.5 diffusion · UDC unknown | PSNR: 25.57 -> 23.05 | Hurt |
| add iterative conditioning during training (ablation) | SD3.5 flow matching · UDC unknown | PSNR: 28.49 -> 28.60 | Improved |
| use RAR stopping criterion (2.4 rounds on average) vs fixed 4 rounds (ablation) | RAR · UDC unknown | PSNR: 28.60 -> 29.69 | Improved |
| use RAR iterative restoration vs AgenticIR | AgenticIR · Composite Group C | wall clock time (s): 78.00 -> 6.92 | Improved |
| use RAR vs AutoDIR on single-degradation tasks | AutoDIR · Single degradation average | PSNR: 27.81 -> 25.88 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 0 of 1 on most shared results.
- AutoDIR — AutoDIR: Automatic All-in-One Image Restoration with Latent Diffusion (ECCV 2024) · better on 1 of 2 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- UARE: A Unified Vision-Language Model for Image Quality Assessment, Restoration, and Enhancement (CVPR 2026 · similarity 0.91)
- Continuous Expert Assembly: Instance-Conditioned Low-Rank Residuals for All-in-One Image Restoration (NeurIPS 2026 · similarity 0.89)
- Learning Where and What to Restore for Composite Image Restoration (NeurIPS 2026 · similarity 0.89)
- DreamClear: High-Capacity Real-World Image Restoration with Privacy-Safe Dataset Curation (NeurIPS 2024 · similarity 0.89)
- From Zero to Detail: Deconstructing Ultra-High-Definition Image Restoration from Progressive Spectral Perspective (CVPR 2025 · similarity 0.88)
- Test-Time Degradation Adaptation for Open-Set Image Restoration (ICML 2024 · similarity 0.88)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
