Research map / Image generation / Image restoration
Patch2Self: Denoising Diffusion MRI with Self-Supervised Learning
Shreyas Fadnavis, Joshua Batson, Eleftherios Garyfallidis
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 Patch2Self instead of Marchenko-Pastur | DWI denoising · Simulated DWI, SNR 5–30 | R²: 0.73 -> 0.84 (SNR 15) | Improved |
| Use Patch2Self instead of Marchenko-Pastur | DWI denoising · Simulated DWI, SNR 5–25 | RMSE: 2.79 -> 2.74 (SNR 15) | Improved |
| Use Patch2Self instead of Marchenko-Pastur | DWI denoising · Simulated DWI, SNR 30 | RMSE: 0.98 -> 0.98 | No clear effect |
| Use Patch2Self instead of Marchenko-Pastur | DWI denoising · Corpus callosum and centrum semiovale voxels | R²: 22%; 65% | Improved |
| Use Patch2Self instead of Marchenko-Pastur | Probabilistic optic-radiation tractography · Optic Radiation bundle | Fiber Bundle Coherency | Improved |
| Increase patch radius from zero to one (ablation) | Patch2Self · Simulated DWI at SNR 15 | RMSE | Mixed |
| Use a simple linear regressor rather than more sophisticated regressors (ablation) | Patch2Self · DWI denoising | Denoising accuracy | No clear effect |
Later papers that compared with this method
2 later accepted papers report Patch2Self as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- Uncertainty-Propelled Physics-MAE Fusion for Self-Supervised Diffusion-Weighted Image Denoising (AAAI 2026 · better on 5 of 5 shared results)
- Self-Supervised Diffusion MRI Denoising via Iterative and Stable Refinement (ICLR 2025 · better on 1 of 3 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
- Patch2Self2: Self-supervised Denoising on Coresets via Matrix Sketching (CVPR 2024 · similarity 0.93)
- Self-Supervised Diffusion MRI Denoising via Iterative and Stable Refinement (ICLR 2025 · similarity 0.93)
- Score-based Self-supervised MRI Denoising (ICLR 2025 · similarity 0.90)
- Patched Denoising Diffusion Models For High-Resolution Image Synthesis (ICLR 2024 · similarity 0.87)
- Rethinking Diffusion Model for Multi-Contrast MRI Super-Resolution (CVPR 2024 · similarity 0.86)
- Denoising Diffusion Bridge Models (ICLR 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.
