Research map / Image generation / Image restoration
FBI-Denoiser: Fast Blind Image Denoiser for Poisson-Gaussian Noise
Check 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 |
|---|---|---|---|
| Use FBI-D instead of BP-AIDE | Blind Poisson-Gaussian image denoising · Fivek synthetic, (0.01, 0.02) | PSNR: 42.59 -> 43.14 | Helped |
| Use FBI-D instead of BP-AIDE | Blind Poisson-Gaussian image denoising · FMD CF MICE | PSNR: 38.31 -> 38.32 | No clear effect |
| Use FBI-D instead of BP-AIDE | Blind Poisson-Gaussian image denoising · SIDD RAW | PSNR: 50.45 -> 50.57 | Helped |
| Use FBI-D trained on SIDD instead of BP-AIDE trained on SIDD | Blind Poisson-Gaussian image denoising · DND RAW | PSNR: 47.75 -> 48.02 | Helped |
| Estimate noise parameters with PGE-Net instead of Liu | GAT+BM3D · Fivek synthetic, mixture noise | PSNR: 38.58 -> 38.72 | Helped |
| Replace PGE-Net's stabilized-variance loss with supervised parameter MSE (ablation) | PGE-Net noise estimation · Fivek synthetic, mixture noise | PSNR: 38.72 -> 35.84 | Hurt |
| Add the outer residual connection (ablation) | FBI-Net · Fivek, (0.01, 0.02) | PSNR: 44.11 -> 44.38 | Helped |
| Add the inner residual connection (ablation) | FBI-Net · Fivek, (0.01, 0.02) | PSNR: 44.31 -> 44.38 | Helped |
| Add the residual module (ablation) | FBI-Net · Fivek, (0.01, 0.02) | PSNR: 44.19 -> 44.38 | Helped |
| Add the inner residual connection and residual module (ablation) | FBI-Net with outer RC · Fivek, (0.01, 0.02) | PSNR: 28.82 -> 44.38 | Helped |
| Add the outer residual connection and residual module (ablation) | FBI-Net with inner RC · Fivek, (0.01, 0.02) | PSNR: 28.84 -> 44.38 | Helped |
| Add the outer and inner residual connections (ablation) | FBI-Net with RM · Fivek, (0.01, 0.02) | PSNR: 28.82 -> 44.38 | Helped |
Closest papers by meaning
- Convexity-Aware Noise Calibration: A Self-Supervised Framework for Noise-Level-Unknown Image Denoising (CVPR 2026 · similarity 0.89)
- Toward Convolutional Blind Denoising of Real Photographs (CVPR 2019 · similarity 0.89)
- Blind2Unblind: Self-Supervised Image Denoising With Visible Blind Spots (CVPR 2022 · similarity 0.89)
- APR-RD: Complemental Two Steps for Self-Supervised Real Image Denoising (AAAI 2025 · similarity 0.89)
- Zero-Shot Blind-spot Image Denoising via Implicit Neural Sampling (CVPR 2025 · similarity 0.87)
- Masked and Shuffled Blind Spot Denoising for Real-World Images (CVPR 2024 · similarity 0.87)
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.