Research map / Image generation / Image restoration
Image Restoration via Diffusion Models with Dynamic Resolution
Yang Zheng, Wen Li, Zhaoqiang Liu
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 |
|---|---|---|---|
| add threshold tau to stop noise injection late in sampling (tau 3e-4 vs 0) (ablation) | SubDAPS++ · FFHQ Gaussian deblurring | FID: 65.14 -> 59.11 | Improved |
| raise noise-suppression threshold tau (ablation) | SubDAPS++ · FFHQ Gaussian deblurring | PSNR: 29.42 -> 28.95 | Hurt |
| add corrector step (ablation) | SubDAPS++ · FFHQ high dynamic range | PSNR: 25.05 -> 25.52 | Improved |
| add corrector step (ablation) | SubDAPS++ · ImageNet high dynamic range | PSNR: 23.57 -> 24.30 | Improved |
| use fine-tuned dynamic-resolution prior vs off-the-shelf Subspace DM prior (ablation) | SubDAPS++ · FFHQ super-resolution 4x | FID: 119.78 -> 64.61 | Improved |
| use ensemble of resolution-specific models instead of fine-tuned model (ablation) | SubDAPS++ · FFHQ super-resolution 4x | PSNR: 29.41 vs 29.34 | No clear effect |
| adapt DAPS to dynamic resolution (DAPS -> SubDAPS) | DAPS · FFHQ inpainting and others | PSNR: 28.88 -> 28.89 (SR 4x) | No clear effect |
| adapt DPS to dynamic resolution (DPS -> SubDPS) | DPS · FFHQ super-resolution 4x | PSNR: 24.00 -> 23.74 | Hurt |
| run SubDAPS++ at full dimension (SubDAPS-F++) instead of dynamic resolution (ablation) | SubDAPS++ · six restoration tasks | PSNR/FID | No clear effect |
| use dynamic resolution (SubDAPS++ vs SubDAPS-F++) for speed (ablation) | SubDAPS++ · ImageNet high dynamic range | time (s): 9.2 -> 7.4 | Improved |
| operate SubDPS at 64x64 resolution vs DPS | SubDPS · ImageNet HDR | memory | Improved |
| add SubDAPS++ improvements (CG, late deterministic, corrector) on SubDAPS | SubDAPS · FFHQ motion deblurring | FID: 82.45 -> 60.43 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- FlowDPS — FlowDPS : Flow-Driven Posterior Sampling for Inverse Problems (ICCV 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
- DGSolver: Diffusion Generalist Solver with Universal Posterior Sampling for Image Restoration (NeurIPS 2025 · similarity 0.91)
- A Variational Perspective on Solving Inverse Problems with Diffusion Models (ICLR 2024 · similarity 0.90)
- A Modular Conditional Diffusion Framework for Image Reconstruction (NeurIPS 2024 · similarity 0.90)
- Ambient Diffusion Posterior Sampling: Solving Inverse Problems with Diffusion Models Trained on Corrupted Data (ICLR 2025 · similarity 0.90)
- Improving Diffusion-Based Image Restoration with Error Contraction and Error Correction (AAAI 2024 · similarity 0.90)
- Taming Diffusion Prior for Image Super-Resolution with Domain Shift SDEs (NeurIPS 2024 · similarity 0.90)
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.
