Research map / Image generation / Image restoration
Diffusion Once and Done: Degradation-Aware LoRA for All-in-One Image Restoration
Ni Tang, Xiaotong Luo, Zihan Cheng, Liangtai Zhou, Dongxiao Zhang, Yanyun Qu
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 conditional LoRA (CLoRA) (ablation) | DOD · Dehazing/deraining/denoising average (ablation) | MUSIQ: 64.81 -> 67.00 | Improved |
| add conditional LoRA (CLoRA) (ablation) | DOD · Dehazing/deraining/denoising average (ablation) | DISTS: 0.1121 -> 0.0873 | Improved |
| add detail enhancement module (DEM) (ablation) | DOD · Dehazing/deraining/denoising average (ablation) | PSNR (dB): 23.05 -> 31.83 | Improved |
| add CLoRA on top of DEM (ablation) | DOD · Dehazing/deraining/denoising average (ablation) | MUSIQ: 68.30 -> 68.85 | Improved |
| add CLoRA on top of DEM (ablation) | DOD · Dehazing/deraining/denoising average (ablation) | PSNR (dB): 31.83 -> 31.87 | No clear effect |
| apply CLoRA also in the VAE encoder (ablation) | DOD · Dehazing/deraining/denoising average (ablation) | PSNR (dB): 31.87 -> 30.68 | Hurt |
| raise LoRA rank from 8 to 16 (ablation) | DOD · Dehazing/deraining/denoising average (ablation) | MUSIQ: 68.85 -> 68.13 | Hurt |
| lower LoRA rank to 4 (ablation) | DOD · Dehazing/deraining/denoising average (ablation) | all metrics | Hurt |
| use one-step diffusion with DMD instead of multi-step sampling | DOD · 512x512 inference | time (s): 17.664 -> 0.211 | Improved |
| use SD-based DOD instead of fidelity-focused non-diffusion methods | DOD · Dehazing/deraining/denoising average | PSNR (dB): 32.98 -> 31.87 | Hurt |
| use DOD vs DA-CLIP (diffusion) | DOD · Dehazing/deraining/denoising average | PSNR (dB): 30.27 -> 31.87 | Improved |
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.
- AdaIR — AdaIR: Adaptive All-in-One Image Restoration via Frequency Mining and Modulation (ICLR 2025) · better on 0 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
- Diffusion Models for Image Restoration and Enhancement: A Comprehensive Survey (IJCV 2025 · similarity 0.91)
- Learning Diffusion Texture Priors for Image Restoration (CVPR 2024 · similarity 0.90)
- GenDeg: Diffusion-based Degradation Synthesis for Generalizable All-In-One Image Restoration (CVPR 2025 · similarity 0.90)
- Visual-Instructed Degradation Diffusion for All-in-One Image Restoration (CVPR 2025 · similarity 0.89)
- UP-Restorer: When Unrolling Meets Prompts for Unified Image Restoration (AAAI 2025 · similarity 0.89)
- RAP-SR: RestorAtion Prior Enhancement in Diffusion Models for Realistic Image Super-Resolution (AAAI 2025 · similarity 0.89)
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.
