Research map / Image generation / Image restoration
Spatial-Adaptive Network for Single Image Denoising
Meng Chang, Qi Li, Huajun Feng, Zhihai Xu
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 |
|---|---|---|---|
| add RSAB (modulated deformable conv blocks) in place of ResBlock (ablation) | SADNet baseline · Kodak24 color (sigma=50, ablation) | PSNR: 29.05 -> 29.55 | Improved |
| add context block (ablation) | SADNet baseline · Kodak24 color (sigma=50, ablation) | PSNR: 29.05 -> 29.12 | Improved |
| add context block to RSAB (ablation) | SADNet (RSAB) · Kodak24 color (sigma=50, ablation) | PSNR: 29.55 -> 29.60 | Improved |
| add offset transfer to RSAB (ablation) | SADNet (RSAB) · Kodak24 color (sigma=50, ablation) | PSNR: 29.55 -> 29.59 | Improved |
| add offset transfer to RSAB + context block (ablation) | SADNet · Kodak24 color (sigma=50, ablation) | PSNR: 29.60 -> 29.64 | Improved |
| add context block to RSAB + offset transfer (ablation) | SADNet · Kodak24 color (sigma=50, ablation) | PSNR: 29.59 -> 29.64 | Improved |
| use dilation rates 1,2,4,8 instead of 1,2,3,4 in context block (ablation) | SADNet · DND sRGB | PSNR / SSIM: 39.59 -> 39.37 PSNR; 0.9523 -> 0.9544 SSIM | Mixed |
| SADNet versus RNAN on grayscale Kodak24 sigma=30 | SADNet · Kodak24 grayscale (sigma=30) | PSNR: 30.04 -> 30.00 | Hurt |
| SADNet versus RIDNet on SIDD validation | SADNet · SIDD validation | PSNR: 38.71 -> 39.46 | Improved |
| SADNet versus RIDNet on Nam | SADNet · Nam (JPEG) | PSNR: 41.04 -> 42.92 | Improved |
| SADNet versus RNAN on efficiency | SADNet · 480x320 color images | FLOPs: 1163.5G -> 50.1G | Improved |
Closest papers by meaning
- Robust And Interpretable Blind Image Denoising Via Bias-Free Convolutional Neural Networks (ICLR 2020 · similarity 0.90)
- Real Image Denoising With Feature Attention (ICCV 2019 · similarity 0.90)
- Generative Adaptive Convolutions for Real-World Noisy Image Denoising (AAAI 2022 · similarity 0.89)
- Spatially Adaptive Self-Supervised Learning for Real-World Image Denoising (CVPR 2023 · similarity 0.89)
- APR-RD: Complemental Two Steps for Self-Supervised Real Image Denoising (AAAI 2025 · similarity 0.88)
- LAN: Learning to Adapt Noise for Image Denoising (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.
