Research map / Image generation / Image restoration
A Unified Conditional Framework for Diffusion-based Image Restoration
Yi Zhang, Xiaoyu Shi, Dasong Li, Xiaogang Wang, Jian Wang, Hongsheng Li
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 |
|---|---|---|---|
| Use the unified conditional diffusion framework instead of the strongest listed baseline | Image restoration · SID | LPIPS: 0.338 -> 0.222 | Improved |
| Use the unified conditional diffusion framework instead of the strongest listed baseline | Image restoration · GoPro | LPIPS: 0.084 -> 0.080 | Improved |
| Use the unified conditional diffusion framework instead of the strongest listed baseline | Image restoration · ImageNet JPEG restoration, Quality Factor 10 | LPIPS: 0.205 -> 0.139 | Improved |
| Add LayerNorm to the diffusion block (ablation) | Conditional diffusion framework · SID | LPIPS: 0.227 -> 0.222 | Improved |
| Use LayerNorm rather than GroupNorm (ablation) | Conditional diffusion framework · SID | LPIPS: 0.240 -> 0.222 | Improved |
| Use Swish rather than ReLU (ablation) | Conditional diffusion framework · SID | LPIPS: 0.235 -> 0.222 | Improved |
| Add initial-predictor spatial guidance (ablation) | Conditional diffusion framework · SID | LPIPS: 0.240 -> 0.222 | Improved |
| Use initial-predictor output rather than internal features as spatial guidance (ablation) | Conditional diffusion framework · SID | LPIPS: 0.247 -> 0.222 | Improved |
| Use initial-predictor output rather than the degraded image as spatial guidance (ablation) | Conditional diffusion framework · SID | LPIPS: 0.253 -> 0.222 | Improved |
| Use adaptive kernel integration rather than direct addition (ablation) | Conditional diffusion framework · SID | LPIPS: 0.236 -> 0.222 | Improved |
| Use adaptive kernel integration rather than concatenation (ablation) | Conditional diffusion framework · SID | LPIPS: 0.241 -> 0.222 | Improved |
| Use adaptive kernel integration rather than AdaIN (ablation) | Conditional diffusion framework · SID | LPIPS: 0.240 -> 0.222 | 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.
- Restormer — Restormer: Efficient Transformer for High-Resolution Image Restoration (CVPR 2022) · better on 2 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
- A Modular Conditional Diffusion Framework for Image Reconstruction (NeurIPS 2024 · similarity 0.93)
- Residual Denoising Diffusion Models (CVPR 2024 · similarity 0.92)
- Learning Diffusion Texture Priors for Image Restoration (CVPR 2024 · similarity 0.91)
- A Simple Approach to Unifying Diffusion-based Conditional Generation (ICLR 2025 · similarity 0.91)
- Improving Diffusion-Based Image Restoration with Error Contraction and Error Correction (AAAI 2024 · similarity 0.91)
- DEFT: Efficient Fine-tuning of Diffusion Models by Learning the Generalised $h$-transform (NeurIPS 2024 · similarity 0.91)
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.
