Research map / Image generation / Image restoration
Toward Convolutional Blind Denoising of Real Photographs
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 the full CBDNet with synthetic and real training images | CBDNet · DND | PSNR: 37.94 -> 38.06 | Helped |
| use the full CBDNet with synthetic and real training images | CBDNet · DND | SSIM: 0.9403 -> 0.9421 | Helped |
| use JPEG-aware training (ablation) | CBDNet · Nam | PSNR: 40.02 -> 41.31 | Helped |
| use JPEG-aware training (ablation) | CBDNet · Nam | SSIM: 0.9687 -> 0.9784 | Helped |
| replace Gaussian noise with heterogeneous Gaussian noise without ISP (ablation) | CBDNet · DND | PSNR: 32.52 -> 33.70 | Helped |
| add ISP to the Gaussian noise model (ablation) | CBDNet · DND | PSNR: 32.52 -> 37.41 | Helped |
| add ISP to the heterogeneous Gaussian noise model (ablation) | CBDNet · DND | PSNR: 33.70 -> 37.57 | Helped |
| replace Gaussian noise with heterogeneous Gaussian noise without ISP (ablation) | CBDNet · Nam | PSNR: 37.62 -> 38.40 | Helped |
| add ISP to the Gaussian noise model (ablation) | CBDNet · Nam | PSNR: 37.62 -> 39.03 | Helped |
| add ISP to the heterogeneous Gaussian noise model (ablation) | CBDNet · Nam | PSNR: 38.40 -> 39.20 | Helped |
| include JPEG compression in the ISP noise model (ablation) | CBDNet · Nam | PSNR: 39.20 -> 40.51 | Helped |
| combine synthetic and real training images rather than synthetic images only (ablation) | CBDNet · DND | PSNR: 37.57 -> 38.06 | Helped |
Closest papers by meaning
- Robust And Interpretable Blind Image Denoising Via Bias-Free Convolutional Neural Networks (ICLR 2020 · similarity 0.92)
- Universal Denoising Networks : A Novel CNN Architecture for Image Denoising (CVPR 2018 · similarity 0.92)
- An Unsupervised Deep Learning Approach for Real-World Image Denoising (ICLR 2021 · similarity 0.91)
- Generative Adaptive Convolutions for Real-World Noisy Image Denoising (AAAI 2022 · similarity 0.91)
- LAN: Learning to Adapt Noise for Image Denoising (CVPR 2024 · similarity 0.89)
- sRGB Real Noise Modeling via Noise-Aware Sampling with Normalizing Flows (ICLR 2024 · similarity 0.88)
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.