Research map / Image generation / Image restoration
UniH3: Unifying Hierarchical Homogeneity and Heterogeneity for All-in-One Medical Image Restoration
Zhiwen Yang, Jiayin Li, Chengyu Liu, Hui Zhang, Bingzheng Wei, Yan Xu
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 Hierarchical Homogeneity Memory (H2M) (ablation) | UniH3 baseline · MedIR-2D-500K all-in-one (7 tasks avg) | PSNR: 36.52 -> 36.66 | Improved |
| add Hierarchical Heterogeneity Balancer (H2B) (ablation) | UniH3 baseline · MedIR-2D-500K all-in-one (7 tasks avg) | PSNR: 36.52 -> 36.64 | Improved |
| combine H2M and H2B (ablation) | UniH3 baseline · MedIR-2D-500K all-in-one (7 tasks avg) | PSNR: 36.52 -> 36.77 | Improved |
| use HGA instead of SFT guidance (ablation) | UniH3 · MedIR-2D-500K all-in-one (7 tasks avg) | PSNR: 36.74 -> 36.77 | Improved |
| use HGA instead of cross attention guidance (ablation) | UniH3 · MedIR-2D-500K all-in-one (7 tasks avg) | PSNR: 36.67 -> 36.77 | Improved |
| add intra-task homogeneity prior on top of inter-task (ablation) | H2M · MedIR-2D-500K all-in-one (7 tasks avg) | PSNR: 36.61 -> 36.77 | Improved |
| add sample-level (intra-task) uncertainty to task-level (ablation) | H2B · MedIR-2D-500K all-in-one (7 tasks avg) | PSNR: 36.70 -> 36.77 | Improved |
| apply H2M and H2B to AdaIR (ablation) | AdaIR · MedIR-2D-500K all-in-one (7 tasks avg) | PSNR: 36.56 -> 36.75 | Improved |
| apply H2M and H2B to Restormer (ablation) | Restormer · MedIR-2D-500K all-in-one (7 tasks avg) | PSNR: 36.49 -> 36.67 | Improved |
| all-in-one UniH3 vs single-task MambaIR | UniH3 · MedIR-2D-500K | PSNR: 36.77 vs 36.75 | No clear effect |
| UniH3 vs second-best on 3D single task | UniH3-3D · MedIR-3D-3K single-task | PSNR: 43.81 -> 45.29 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 4 of 4 on most shared results.
- AdaIR — AdaIR: Adaptive All-in-One Image Restoration via Frequency Mining and Modulation (ICLR 2025) · better on 3 of 3 shared results
- MambaIR — MambaIR: A Simple Baseline for Image Restoration with State-Space Model (ECCV 2024) · better on 2 of 2 shared results
- Restormer — Restormer: Efficient Transformer for High-Resolution Image Restoration (CVPR 2022) · better on 2 of 2 shared results
- PromptIR — PromptIR: Prompting for All-in-One Image Restoration (NeurIPS 2023) · 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
- Beyond the Static-World: Lifelong Learning for All-in-One Medical Image Restoration (CVPR 2026 · similarity 0.89)
- Debiased All-in-one Image Restoration with Task Uncertainty Regularization (AAAI 2025 · similarity 0.88)
- UARE: A Unified Vision-Language Model for Image Quality Assessment, Restoration, and Enhancement (CVPR 2026 · similarity 0.87)
- Parameter Efficient Adaptation for Image Restoration with Heterogeneous Mixture-of-Experts (NeurIPS 2024 · similarity 0.86)
- InstructRestore: Region-Customized Image Restoration with Human Instructions (NeurIPS 2025 · similarity 0.86)
- Scaling Up to Excellence: Practicing Model Scaling for Photo-Realistic Image Restoration In the Wild (CVPR 2024 · similarity 0.86)
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.
