Research map / Image generation / Image restoration
Multi-Channel Weighted Nuclear Norm Minimization for Real Color Image Denoising
Jun Xu, Lei Zhang, David Zhang, Xiangchu Feng
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 per-channel noise weight matrix W to concatenated WNNM (MC-WNNM vs WNNM-3) (ablation) | WNNM on concatenated RGB patches · Kodak 24 images | average PSNR (dB): 28.22 -> 29.31 | Improved |
| use concatenated RGB patches with weight matrix instead of per-channel WNNM-1 (ablation) | WNNM · Kodak 24 images | average PSNR (dB): 28.84 -> 29.31 | Improved |
| use weight matrix instead of single averaged noise level in concatenated WNNM-2 (ablation) | WNNM · Kodak 24 images | average PSNR (dB): 28.83 -> 29.31 | Improved |
| add per-channel noise weight matrix W to concatenated WNNM (MC-WNNM vs WNNM-3) (ablation) | WNNM on concatenated RGB patches · real noisy images (15 crops) | average PSNR (dB): 37.12 -> 37.71 | Improved |
| use weight matrix instead of single averaged noise level in concatenated WNNM-2 (ablation) | WNNM · real noisy images (15 crops) | average PSNR (dB): 37.27 -> 37.71 | Improved |
| use concatenated RGB patches with weight matrix instead of per-channel WNNM-1 (ablation) | WNNM · real noisy images (15 crops) | average PSNR (dB): 35.77 -> 37.71 | Improved |
| MC-WNNM instead of WNNM-3 on the simplest (not highest PSNR) real image subsets (ablation) | WNNM · real noisy images (Nikon D800 ISO=3200 and 6400 images) | PSNR (dB): 38.68 vs 39.39 and 41.52 vs 39.39 | Mixed |
| use MC-WNNM instead of Neat Image | denoising · real noisy images (15 crops) | average time (s): 0.9 -> 202.9 | Hurt |
Closest papers by meaning
- Weighted Nuclear Norm Minimization and Its Applications to Low Level Vision (IJCV 2016 · similarity 0.88)
- Multi-Scale Weighted Nuclear Norm Image Restoration (CVPR 2018 · similarity 0.87)
- A Trilateral Weighted Sparse Coding Scheme for Real-World Image Denoising (ECCV 2018 · similarity 0.83)
- C2N: Practical Generative Noise Modeling for Real-World Denoising (ICCV 2021 · similarity 0.82)
- URWKV: Unified RWKV Model with Multi-state Perspective for Low-light Image Restoration (CVPR 2025 · similarity 0.80)
- DnLUT: Ultra-Efficient Color Image Denoising via Channel-Aware Lookup Tables (CVPR 2025 · similarity 0.80)
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.
