Research map / Image generation / Image restoration
Asymmetric Mask Scheme for Self-Supervised Real Image Denoising
Xiangyu Liao, Tianheng Zheng, Jiayu Zhong, Pingping Zhang, Chao Ren
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 random replacement refinement at inference (ablation) | AMSNet-P · SIDD Validation | PSNR (dB): 36.71 -> 37.93 | Improved |
| add smoothness loss fine-tuning L_t (ablation) | AMSNet-B · SIDD Validation | PSNR (dB): 36.60 -> 36.71 | Improved |
| use mask-based training instead of blind-spot with unrestricted receptive field (ablation) | AMSNet vs AP-BSN · SIDD Validation | PSNR (dB): 20.91 -> 37.11 | Improved |
| use Restormer instead of DnCNN as denoiser in AMSNet (ablation) | AMSNet · SIDD Validation | PSNR (dB): 36.93 -> 37.93 | Improved |
| set mask ratio per branch to about 50% | AMSNet · SIDD Validation | PSNR (dB) | Improved |
| use AMSNet-B without refinement versus AP-BSN+R3 | AMSNet-B · SIDD Benchmark | PSNR (dB): 36.60 vs 36.91 | No clear effect |
| use AMSNet-P-E instead of best prior self-supervised methods | AMSNet-P-E · SIDD Benchmark | PSNR (dB): 37.41 -> 37.87 vs BNN-LAN | Improved |
| use AMSNet-P-E on PolyU | AMSNet-P-E · PolyU | PSNR (dB): 37.13 -> 37.92 vs BNN-LAN | Improved |
Closest papers by meaning
- Exploring Efficient Asymmetric Blind-Spots for Self-Supervised Denoising in Real-World Scenarios (CVPR 2024 · similarity 0.94)
- APR-RD: Complemental Two Steps for Self-Supervised Real Image Denoising (AAAI 2025 · similarity 0.90)
- Zero-Shot Noise2Mean: Gap Minimization for Efficient Denoising from a Single Noisy Image (AAAI 2025 · similarity 0.89)
- Rethinking Transformer-Based Blind-Spot Network for Self-Supervised Image Denoising (AAAI 2025 · similarity 0.89)
- Masked and Shuffled Blind Spot Denoising for Real-World Images (CVPR 2024 · similarity 0.88)
- Pseudo-Siamese Blind-spot Transformers for Self-Supervised Real-World Denoising (NeurIPS 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.
