Research map / Image generation / Image restoration
Plug-and-Play Image Restoration With Deep Denoiser Prior
Kai Zhang, Yawei Li, Wangmeng Zuo, Lei Zhang, Luc Van Gool, Radu Timofte
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 skip connections in residual blocks (ablation) | DRUNet · Set12 denoising, sigma=50 | PSNR: 27.78 -> 27.90 | Improved |
| add more training data (ablation) | DRUNet · Set12 denoising, sigma=50 | PSNR: 27.87 -> 27.90 | Improved |
| remove biases (bias-free network) (ablation) | DRUNet · Set12 denoising, sigma=50 | PSNR: 27.89 -> 27.90 | No clear effect |
| remove biases (bias-free network) (ablation) | DRUNet · unseen noise level 200 | visual quality | Improved |
| add noise level map as input (ablation) | DRUNet · Set12 denoising, sigma=50 | PSNR: 27.83 -> 27.90 | Improved |
| add more training data (ablation) | DRUNet · JPEG deblocking | PSNR: 0.04dB average decrease with less data | Improved |
| retrain DnCNN3 with our training data | DnCNN3 · Classic5 and LIVE1 JPEG deblocking | PSNR: +0.21dB and +0.19dB | Improved |
| use non-blind denoiser with noise level map instead of blind denoiser (ablation) | DPIR · image deblurring (Leaves) | PSNR | Improved |
| replace IRCNN denoiser with DRUNet denoiser at same iterations | DPIR · Set6 deblurring | PSNR: 0.2dB~2dB over IRCNN | Improved |
| increase initial noise level sigma_1 from 9 to 49 (ablation) | DPIR · deblurring test image, K=8 | PSNR: 22.50 -> 30.27 | Improved |
| increase iterations K from 8 to 40 (ablation) | DPIR · deblurring test image, sigma_1=49 | PSNR: 30.27 -> 30.01 | No clear effect |
| use stronger DRUNet denoiser in place of IRCNN | DPIR · color demosaicing | PSNR: up to 1.8dB over IRCNN+ | Improved |
Closest papers by meaning
- Denoising Prior Driven Deep Neural Network for Image Restoration (TPAMI 2018 · similarity 0.92)
- Plug-and-Play image restoration with Stochastic deNOising REgularization (ICML 2024 · similarity 0.89)
- A Restoration Network as an Implicit Prior (ICLR 2024 · similarity 0.89)
- Reconciling Stochastic and Deterministic Strategies for Zero-shot Image Restoration using Diffusion Model in Dual (CVPR 2025 · similarity 0.89)
- Image Restoration by Denoising Diffusion Models with Iteratively Preconditioned Guidance (CVPR 2024 · similarity 0.88)
- SP$^3$: Spherical Priors for Plug-and-Play Restoration (NeurIPS 2026 · 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.
