Research map / Image generation / Image restoration
High-Quality Self-Supervised Deep Image Denoising
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Train a blind-spot model with a learned clean-signal prior and use posterior-mean estimation | Image denoising · Gaussian σ=25 | PSNR: 31.60 -> 31.57 | No clear effect |
| Train the self-supervised model and estimate the unknown noise level | Image denoising · Gaussian σ∈[5,50] | PSNR: 31.71 -> 31.59 | Hurt |
| Use the self-supervised Bayesian method instead of N2C | Image denoising · Poisson λ=30 | PSNR: 30.89 -> 30.78 | Hurt |
| Estimate variable Poisson noise parameters from the data | Image denoising · Poisson λ∈[5,50] | PSNR: 30.40 -> 29.79 | Hurt |
| Use the self-supervised Bayesian method instead of N2C | Image denoising · Impulse α=0.5 | PSNR: 31.98 -> 31.57 | Hurt |
| Use the self-supervised Bayesian method instead of N2C | Image denoising · Impulse α∈[0,1] | PSNR: 30.58 -> 30.29 | Hurt |
| Allow a full covariance matrix instead of forcing diagonal covariance (ablation) | Self-supervised Bayesian denoising · Gaussian σ=25 | PSNR: 30.70 -> 31.57 | Helped |
| Use posterior-mean estimation to incorporate the noisy center pixel (ablation) | Self-supervised Bayesian denoising · Gaussian σ=25 | PSNR: 29.62 -> 31.57 | Helped |
| Use posterior-mean estimation to incorporate the noisy center pixel (ablation) | Self-supervised Bayesian denoising · Poisson λ=30 | PSNR: 29.17 -> 30.78 | Helped |
| Use posterior-mean estimation to incorporate the noisy center pixel (ablation) | Self-supervised Bayesian denoising · Impulse α∈[0,1] | PSNR: 26.09 -> 30.29 | Helped |
| Add rotation augmentation | Self-supervised Bayesian denoising · Gaussian σ=25, 100 training images | PSNR: 31.03 -> 31.08 | Helped |
Closest papers by meaning
- Convexity-Aware Noise Calibration: A Self-Supervised Framework for Noise-Level-Unknown Image Denoising (CVPR 2026 · similarity 0.92)
- Blind2Unblind: Self-Supervised Image Denoising With Visible Blind Spots (CVPR 2022 · similarity 0.92)
- Robust And Interpretable Blind Image Denoising Via Bias-Free Convolutional Neural Networks (ICLR 2020 · similarity 0.91)
- APR-RD: Complemental Two Steps for Self-Supervised Real Image Denoising (AAAI 2025 · similarity 0.90)
- Pseudo-Siamese Blind-spot Transformers for Self-Supervised Real-World Denoising (NeurIPS 2024 · similarity 0.89)
- Zero-Shot Noise2Mean: Gap Minimization for Efficient Denoising from a Single Noisy Image (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.