Research map / Detection and segmentation / CNN features
Deep Semantic Statistics Matching (D2SM) Denoising Network
Kangfu Mei, Vishal M. Patel, Rui Huang
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 internal patch-wise probabilities (ablation) | D2SM on FFDNet · Cityscapes denoising (sigma=25) | PSNR: 35.543 -> 36.454 | Improved |
| add internal patch-wise probabilities (ablation) | D2SM on FFDNet · Cityscapes denoising (sigma=25) | MIoU: 0.612 -> 0.644 | Improved |
| add internal patch-wise probabilities (ablation) | D2SM on CBDNet · Cityscapes denoising (sigma=25) | PSNR: 36.254 -> 36.899 | Improved |
| add internal patch-wise probabilities (ablation) | D2SM on SADNet · Cityscapes denoising (sigma=25) | PSNR: 36.822 -> 37.130 | Improved |
| add D2SM objective to original loss | CBDNet · Cityscapes denoising (sigma=25) | PSNR: 36.152 -> 36.899 | Improved |
| use D2SM objective instead of cross-entropy with segmentation labels | FFDNet · Cityscapes denoising (sigma=25) | PSNR: 35.913 -> 36.454 | Improved |
| add contextual loss to L1 | FFDNet · Cityscapes denoising (sigma=25) | PSNR: 35.033 -> 25.115 | Hurt |
| add perceptual loss to L1 | FFDNet · Cityscapes denoising (sigma=25) | PSNR: 35.033 -> 34.319 | Hurt |
| use KL divergence instead of Jensen-Shannon divergence (ablation) | D2SM on FFDNet · Cityscapes denoising (sigma=25) | PSNR: 36.31 -> 36.45 | Improved |
| use KL divergence instead of inverse KL divergence (ablation) | D2SM on FFDNet · Cityscapes denoising (sigma=25) | PSNR: 35.97 -> 36.45 | Improved |
| use KL divergence instead of GAN objective (ablation) | D2SM on FFDNet · Cityscapes denoising (sigma=25) | PSNR: 35.55 -> 36.45 | Improved |
| add D2SM objective to MSBDN-DFF dehazing | MSBDN-DFF · I-HAZE | PSNR: 23.93 -> 24.31 | Improved |
Closest papers by meaning
- Noise2Score: Tweedie’s Approach to Self-Supervised Image Denoising without Clean Images (NeurIPS 2021 · similarity 0.88)
- Universal Denoising Networks : A Novel CNN Architecture for Image Denoising (CVPR 2018 · similarity 0.88)
- Unpaired Learning of Deep Image Denoising (ECCV 2020 · similarity 0.87)
- Unsupervised Image Restoration Using Partially Linear Denoisers (TPAMI 2021 · similarity 0.87)
- Convexity-Aware Noise Calibration: A Self-Supervised Framework for Noise-Level-Unknown Image Denoising (CVPR 2026 · similarity 0.86)
- Self-Supervised Image Denoising Using Implicit Deep Denoiser Prior (AAAI 2023 · 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.
