Research map / Image generation / Image restoration
Deep Denoising of Flash and No-Flash Pairs for Photography in Low-Light Environments
Zhihao Xia, Michael Gharbi, Federico Perazzi, Kalyan Sunkavalli, Ayan Chakrabarti
Check 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 |
|---|---|---|---|
| use the flash/no-flash filtering-and-scale architecture with large kernels instead of BPN on a flash/no-flash pair | Flash/no-flash denoising · 100x Dimmed | PSNR: 26.11 dB -> 26.75 dB | Improved |
| use the flash/no-flash filtering-and-scale architecture with large kernels instead of KPN on a flash/no-flash pair | Flash/no-flash denoising · 100x Dimmed | SSIM: 0.815 -> 0.829 | Improved |
| use the flash/no-flash filtering-and-scale architecture with large kernels instead of BPN on a flash/no-flash pair | Flash/no-flash denoising · 50x Dimmed | PSNR: 28.04 dB -> 28.56 dB | Improved |
| use the flash/no-flash filtering-and-scale architecture with large kernels instead of KPN on a flash/no-flash pair | Flash/no-flash denoising · 50x Dimmed | SSIM: 0.852 -> 0.860 | Improved |
| use the flash/no-flash filtering-and-scale architecture with large kernels instead of BPN on a flash/no-flash pair | Flash/no-flash denoising · 25x Dimmed | PSNR: 29.75 dB -> 30.14 dB | Improved |
| use the flash/no-flash filtering-and-scale architecture with large kernels instead of KPN on a flash/no-flash pair | Flash/no-flash denoising · 25x Dimmed | SSIM: 0.880 -> 0.884 | Improved |
| use the flash/no-flash filtering-and-scale architecture with large kernels instead of BPN on a flash/no-flash pair | Flash/no-flash denoising · 12.5x Dimmed | PSNR: 31.21 dB -> 31.52 dB | Improved |
| use the flash/no-flash filtering-and-scale architecture with large kernels instead of KPN on a flash/no-flash pair | Flash/no-flash denoising · 12.5x Dimmed | SSIM: 0.901 -> 0.903 | Improved |
| add the flash image as geometric reference rather than use the no-flash image (ablation) | Ours · Dimmed image test settings | PSNR: 26.75 dB -> 26.83 dB; 28.56 dB -> 28.39 dB; 30.14 dB -> 29.55 dB; 31.52 dB -> 30.45 dB | Mixed |
| add the low-frequency interpolated kernel component B_j to use larger kernels (ablation) | Ours · Dimmed image test settings | PSNR: 26.45 dB -> 26.75 dB; 28.42 dB -> 28.56 dB; 30.09 dB -> 30.14 dB; 31.51 dB -> 31.52 dB | Improved |
| use one flash image alongside a no-flash burst rather than an all-no-flash burst | Burst denoising · 100x Dimmed, 5-image burst | PSNR: 25.90 dB -> 26.82 dB | Improved |
| use one flash image alongside a no-flash burst rather than an all-no-flash burst | Burst denoising · 100x Dimmed, 7-image burst | PSNR: 26.00 dB -> 26.80 dB | Improved |
Closest papers by meaning
- Interpretable Unsupervised Joint Denoising and Enhancement for Real-World low-light Scenarios (ICLR 2025 · similarity 0.87)
- Bright-NeRF: Brightening Neural Radiance Field with Color Restoration from Low-Light RAW Images (AAAI 2025 · similarity 0.86)
- LuSh-NeRF: Lighting up and Sharpening NeRFs for Low-light Scenes (NeurIPS 2024 · similarity 0.86)
- ZERO-IG: Zero-Shot Illumination-Guided Joint Denoising and Adaptive Enhancement for Low-Light Images (CVPR 2024 · similarity 0.86)
- APR-RD: Complemental Two Steps for Self-Supervised Real Image Denoising (AAAI 2025 · similarity 0.86)
- Contrastive Learning for Low-Light Raw Denoising (Student Abstract) (AAAI 2024 · similarity 0.86)
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.
