Research map / Image generation / Image restoration
CARD: Correlation Aware Restoration with Diffusion
Niki Nezakati, Arnab Ghosh, Amit Roy-Chowdhury, Vishwanath Saragadam
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 |
|---|---|---|---|
| whiten the observation and use noise-whitened updates instead of DDRM i.i.d. updates | CARD · ImageNet denoising sigma0=0.1 | PSNR: 31.0 -> 34.0 | Improved |
| same, at high noise | CARD · ImageNet denoising sigma0=0.9 | PSNR: 22.7 -> 26.7 | Improved |
| same, on deblurring | CARD · ImageNet Gaussian deblurring sigma0=0.5 | PSNR: 22.2 -> 24.4 | Improved |
| same, on 4x super-resolution | CARD · ImageNet SR4 sigma0=0.5 | PSNR: 20.4 -> 23.1 | Improved |
| same, on real correlated noise | CARD · CIN-D denoising high noise | PSNR: 30.8 -> 31.5 | Improved |
| perturb the covariance by 5% (ablation) | CARD · covariance sensitivity | PSNR: 29.5 -> 26.5 | Hurt |
| perturb the covariance by 20% (ablation) | CARD · covariance sensitivity | PSNR: 29.5 -> 23.6 | Hurt |
| use a covariance estimated from a different camera (Blackfly applied to Nikon Z30) (ablation) | CARD · Z30 images | image quality vs DDRM | No clear effect |
| use 4x128 covariance patch instead of 8x8 (ablation) | CARD · CIN-D indoor high noise | PSNR: 33.1 -> 33.4 | Improved |
| use 32x32 patch instead of 8x8 (ablation) | CARD · CIN-D indoor low noise | PSNR: 39.8 -> 39.2 | Hurt |
Closest papers by meaning
- Residual Denoising Diffusion Models (CVPR 2024 · similarity 0.89)
- Resfusion: Denoising Diffusion Probabilistic Models for Image Restoration Based on Prior Residual Noise (NeurIPS 2024 · similarity 0.89)
- Diffusion Priors for Variational Likelihood Estimation and Image Denoising (NeurIPS 2024 · similarity 0.88)
- How I Warped Your Noise: a Temporally-Correlated Noise Prior for Diffusion Models (ICLR 2024 · similarity 0.88)
- Denoising as Adaptation: Noise-Space Domain Adaptation for Image Restoration (ICLR 2025 · similarity 0.88)
- Improving Diffusion-Based Image Restoration with Error Contraction and Error Correction (AAAI 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.
