Research map / Image generation / Fast diffusion sampling
Efficient Burst Raw Denoising with Variance Stabilization and Multi-frequency Denoising Network
Dasong Li, Yi Zhang, Ka Lung Law, Xiaogang Wang, Hongwei Qin, Hongsheng Li
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 variance-stabilization noise prior integration (ablation) | Ours · CRVD (N=3, raw) | PSNR: 43.99 -> 44.43 | Improved |
| use generalized Freeman-Tukey instead of GAT (ablation) | Ours · CRVD (N=3, raw) | PSNR: 44.34 -> 44.43 | Improved |
| use variance stabilization instead of noise map (ablation) | Ours · CRVD (N=3, raw) | PSNR: 44.26 -> 44.43 | Improved |
| add explicit alignment to KPN* (ablation) | KPN* · KPN synthetic (avg) | PSNR: 32.47 -> 33.09 | Improved |
| add alignment to BPN (ablation) | BPN · KPN synthetic (avg) | PSNR: 33.90 -> 34.22 | Improved |
| train denoiser on aligned images (ablation) | Ours · KPN synthetic (avg) | PSNR: 34.19 -> 34.90 | Improved |
| add alignment at test time only (ablation) | Ours w/o alignment · KPN synthetic (avg) | PSNR: 33.39 -> 34.19 | Improved |
| use deformable convolution alignment instead of conventional alignment (ablation) | Ours · CRVD (N=3, raw) | PSNR: 44.43 -> 44.46 | No clear effect |
| add alignment under large synthetic misalignment (ablation) | Ours · CRVD +-16 px (N=3, raw) | PSNR: 43.68 -> 44.05 | Improved |
| add sequential denoising of alternate frames (ablation) | Ours · CRVD (N=5, raw) | PSNR: 44.37 -> 44.70 | Improved |
| add multi-frequency aggregation (ablation) | Ours · CRVD (N=3, raw) | PSNR: 44.17 -> 44.43 | Improved |
| use specialized weights per sub-network (ablation) | Ours · CRVD (N=5, raw) | PSNR: 44.44 -> 44.70 | Improved |
Closest papers by meaning
- Deep Burst Denoising (ECCV 2018 · similarity 0.90)
- Practical Deep Raw Image Denoising on Mobile Devices (ECCV 2020 · similarity 0.88)
- Burst Denoising via Temporally Shifted Wavelet Transforms (ECCV 2020 · similarity 0.88)
- A Decoupled Learning Scheme for Real-world Burst Denoising from Raw Images (ECCV 2020 · similarity 0.88)
- Real-World Mobile Image Denoising Dataset with Efficient Baselines (CVPR 2024 · similarity 0.85)
- Noise Modeling in One Hour: Minimizing Preparation Efforts for Self-supervised Low-Light RAW Image Denoising (CVPR 2025 · similarity 0.85)
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.
