Research map / Image generation / Image restoration
GRIDS: Grouped Multiple-Degradation Restoration with Image Degradation Similarity
Shuo Cao, Yihao Liu, Wenlong Zhang, Yu Qiao, Chao Dong
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 |
|---|---|---|---|
| group 11 degradations into 4 similarity-based groups instead of mix-training all | Restormer · DIV2K-Valid | average PSNR: 32.98 -> 35.22 | Improved |
| group 11 degradations into 4 similarity-based groups instead of mix-training all | SRResNet · DIV2K-Valid | average PSNR: 31.88 -> 33.32 | Improved |
| mix-train all 11 degradations in a single model | Restormer · DIV2K-Valid | average PSNR: 35.13 -> 32.98 | Hurt |
| group by high DDR similarity vs low similarity (2 to 5 degradations) (ablation) | Mix-training within group · Exp 2, Restormer | avg PSNR drop: -0.71 -> -0.30 | Improved |
| estimate similarity with VGG instead of DDR features (ablation) | Degradation similarity · Exp 5, Restormer | avg PSNR drop: higher-similarity group -1.06 vs lower-similarity group -0.49 | Hurt |
| estimate similarity with SIFT instead of DDR features (ablation) | Degradation similarity · Exp 6, Restormer | avg PSNR drop: higher-similarity group -1.35 vs lower-similarity group -0.36 | Hurt |
| rely on DDR similarity when similarity gap is tiny (ablation) | Degradation similarity · Exp 7 (gap 0.12) | avg PSNR change: 0.05 vs 0.15 | No clear effect |
| add adaptive model selection via GGD KL divergence | GRIDS inference · Real-world noisy/blurry/hazy images | group assignment | Improved |
Closest papers by meaning
- Retrieve-to-Restore: Efficient All-in-One Image Restoration with a Retrieval-Based Degradation Bank (CVPR 2026 · 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 0.89)
- Visual-Instructed Degradation Diffusion for All-in-One Image Restoration (CVPR 2025 · similarity 0.88)
- GenDeg: Diffusion-based Degradation Synthesis for Generalizable All-In-One Image Restoration (CVPR 2025 · similarity 0.88)
- Complexity Experts are Task-Discriminative Learners for Any Image Restoration (CVPR 2025 · 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.
