Research map / Label-efficient and robust learning / Self-supervised and contrastive
Blind2Unblind: Self-Supervised Image Denoising With Visible Blind Spots
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 Blind2Unblind instead of NBR2NBR | raw-RGB denoising · SIDD Benchmark | PSNR (dB): 50.47 -> 50.79 | Helped |
| use Blind2Unblind instead of NBR2NBR | raw-RGB denoising · SIDD Validation | PSNR (dB): 51.06 -> 51.36 | Helped |
| replace L_B with L_A (ablation) | Blind2Unblind · Kodak, Gaussian σ=25 | PSNR (dB): 31.37 -> 32.27 | Helped |
| replace random mask mapper with global-aware mask mapper (ablation) | Blind2Unblind · Kodak, Gaussian σ=25 | PSNR (dB): 30.32 -> 30.56 | Helped |
| combine re-visible loss with global-aware mask mapper (ablation) | Blind2Unblind · Kodak, Gaussian σ=25 | PSNR (dB): 30.56 -> 32.27 | Helped |
| combine re-visible loss with random mask mapper (ablation) | Blind2Unblind · Kodak, Gaussian σ=25 | PSNR (dB): cannot converge | Hurt |
| set the visible-term weight λ_f to 20 rather than 2 (ablation) | Blind2Unblind · Kodak, Gaussian σ=25 | PSNR (dB): 32.11 -> 32.27 | Helped |
| increase λ_f from 20 to 40 (ablation) | Blind2Unblind · Kodak, Gaussian σ=25 | PSNR (dB): 32.27 -> 32.27 | No clear effect |
| remove the regularization term by setting η to 0 (ablation) | Blind2Unblind · Kodak, Gaussian σ=25 | PSNR (dB): 32.19 -> 32.27 | Helped |
| set regularization weight η to 2 rather than 1 (ablation) | Blind2Unblind · Kodak, Gaussian σ=25 | PSNR (dB): 32.27 -> 32.32 | Mixed |
| use Blind2Unblind instead of R2R | grayscale denoising · BSD68, Gaussian σ=50 | PSNR (dB): 26.02 -> 26.09 | Helped |
Closest papers by meaning
- Exploring Efficient Asymmetric Blind-Spots for Self-Supervised Denoising in Real-World Scenarios (CVPR 2024 · similarity 0.92)
- Masked and Shuffled Blind Spot Denoising for Real-World Images (CVPR 2024 · similarity 0.90)
- Rethinking Transformer-Based Blind-Spot Network for Self-Supervised Image Denoising (AAAI 2025 · similarity 0.90)
- Pseudo-Siamese Blind-spot Transformers for Self-Supervised Real-World Denoising (NeurIPS 2024 · similarity 0.90)
- APR-RD: Complemental Two Steps for Self-Supervised Real Image Denoising (AAAI 2025 · similarity 0.89)
- Unsupervised Blind Image Deblurring Based on Self-Enhancement (CVPR 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.