Research map / Image generation / Image restoration
Self-Calibrated Variance-Stabilizing Transformations for Real-World Image Denoising
Sébastien Herbreteau, Michael Unser
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 Noise2VST (learned VST + frozen DRUNet) instead of oracle-parameter GAT baseline | DRUNet · KODAK/BSD300/SET14 Poisson | PSNR: 31.63 -> 31.60 on KODAK; gap less than 0.06 dB at most | No clear effect |
| add learned model-free VST in place of GAT with estimated parameters | DRUNet · SIDD Validation | PSNR: 48.92 -> 51.66 (BM3D+GAT vs Noise2VST) | Improved |
| replace Noise2VST with fast FFDNet-based variant Noise2VST† | Noise2VST · KODAK | PSNR: 31.60 -> 31.51 | No clear effect |
| Noise2VST† instead of Noise2VST at inference | Noise2VST · SIDD Validation | PSNR: 51.66 -> 51.43 | Hurt |
| add Noise2VST over B2UNB | B2UNB · SIDD Benchmark | PSNR: 50.79 -> 51.07 | Improved |
| add Noise2VST over SST-GP | SST-GP · BSD300 | PSNR: 29.96 -> 29.89 | Hurt |
| add Noise2VST over ZS-N2N | ZS-N2N · FMD Confocal Fish | PSNR: 30.62 -> 32.88 | Improved |
| use Noise2VST over supervised N2C baseline | N2C · SIDD Validation | PSNR: 51.19 -> 51.66 | Improved |
| use Noise2VST over supervised N2C baseline | N2C · W2S ch0 avg16 | PSNR: 41.94 -> 39.74 | Hurt |
| Noise2VST over DRUNet + GAT | DRUNet + GAT · W2S ch0 avg1 | PSNR: 32.54 -> 35.65 | Improved |
| Noise2VST over N2V | N2V · FMD Confocal Fish | PSNR: 32.08 -> 32.88 | Improved |
Closest papers by meaning
- Noise2Score: Tweedie’s Approach to Self-Supervised Image Denoising without Clean Images (NeurIPS 2021 · similarity 0.91)
- Unsupervised Image Restoration Using Partially Linear Denoisers (TPAMI 2021 · similarity 0.91)
- Convexity-Aware Noise Calibration: A Self-Supervised Framework for Noise-Level-Unknown Image Denoising (CVPR 2026 · similarity 0.91)
- High-Quality Self-Supervised Deep Image Denoising (NeurIPS 2019 · similarity 0.91)
- Iterative Denoiser and Noise Estimator for Self-Supervised Image Denoising (ICCV 2023 · similarity 0.90)
- Noisier2Noise: Learning to Denoise From Unpaired Noisy Data (CVPR 2020 · 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.
