Research map / Image generation / Image restoration
Retrieve-to-Restore: Efficient All-in-One Image Restoration with a Retrieval-Based Degradation Bank
Chenxu Wang, Kai Zhang, Jian Yang
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 |
|---|---|---|---|
| set degradation bank size M to 64 rather than 24, 32 or 128 (ablation) | R2R · All-in-one 3 degradations (average) | PSNR (dB): 32.30 / 32.41 / 32.34 -> 32.53 | Mixed |
| add FFT frequency loss to pixel loss (ablation) | R2R · All-in-one 3 degradations (average) | PSNR (dB): 31.88 -> 32.23 | Improved |
| add degradation classification loss L_deg (ablation) | R2R · All-in-one 3 degradations (average) | PSNR (dB): 32.23 -> 32.29 | Improved |
| add matching classification loss L_match (ablation) | R2R · All-in-one 3 degradations (average) | PSNR (dB): 32.29 -> 32.53 | Improved |
| build degradation bank from clean HQ and residual instead of Gaussian-noised HQ (ablation) | R2R · All-in-one 3 degradations (average) | PSNR (dB): 31.99 -> 32.53 | Improved |
| build degradation bank from clean HQ and residual instead of LQ only (ablation) | R2R · All-in-one 3 degradations (average) | PSNR (dB): 32.33 -> 32.53 | No clear effect |
| replace PromptIR-style internal modulation with retrieval-based degradation bank | R2R · All-in-one 3 degradations (average) | PSNR (dB) / MACs: 32.06 -> 32.53 PSNR; 132G -> 12G MACs | Improved |
| scale from single-task to three-task training | R2R · single vs three-task restoration | PSNR (dB) | Improved |
| R2R versus PromptIR on single-task denoising | R2R · CBSD68 sigma=15 (single task) | PSNR (dB): 34.34 -> 34.27 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- PromptIR — PromptIR: Prompting for All-in-One Image Restoration (NeurIPS 2023) · better on 7 of 8 shared results
- Up-Restorer — UP-Restorer: When Unrolling Meets Prompts for Unified Image Restoration (AAAI 2025) · better on 1 of 1 shared result
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
- GRIDS: Grouped Multiple-Degradation Restoration with Image Degradation Similarity (ECCV 2024 · similarity 0.91)
- Ingredient-Oriented Multi-Degradation Learning for Image Restoration (CVPR 2023 · similarity 0.91)
- Learning Where and What to Restore for Composite Image Restoration (NeurIPS 2026 · similarity 0.90)
- Continuous Expert Assembly: Instance-Conditioned Low-Rank Residuals for All-in-One Image Restoration (NeurIPS 2026 · similarity 0.90)
- Degradation-Aware Feature Perturbation for All-in-One Image Restoration (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.
