Research map / Image generation / Image restoration
Parallel Diffusion Models of Operator and Image for Blind Inverse Problems
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Use parallel diffusion priors for the image and unknown operator parameters, guided by the measurement likelihood | BlindDPS · FFHQ and AFHQ blind deblurring | FID | Helped |
| Use parallel diffusion priors for the image and unknown operator parameters, guided by the measurement likelihood | BlindDPS · FFHQ and ImageNet turbulence imaging | LPIPS | Helped |
| Use parallel diffusion priors for the image and unknown operator parameters, guided by the measurement likelihood | BlindDPS · FFHQ and ImageNet turbulence imaging | PSNR | Mixed |
| Add a diffusion prior for the blur kernel instead of using a uniform prior (ablation) | BlindDPS · FFHQ blind deblurring (100-image ablation) | LPIPS: 0.566 -> 0.247 | Helped |
| Add a diffusion prior for the blur kernel instead of using a uniform prior (ablation) | BlindDPS · FFHQ blind deblurring (100-image ablation) | PSNR: 11.72 -> 23.65 | Helped |
| Add a diffusion prior for the blur kernel instead of using a uniform prior (ablation) | BlindDPS · FFHQ blind deblurring (100-image ablation) | MSE: 0.163 -> 0.002 | Helped |
| Add kernel sparsity regularization rather than setting λ = 0.0 (ablation) | BlindDPS · FFHQ motion deblurring | MNC: 0.929 -> 0.956 | Helped |
| Increase kernel sparsity weight from λ = 0.0 to λ ≥ 0.1 (ablation) | BlindDPS · FFHQ motion deblurring | MSE: 0.004 -> 0.002 | Helped |
| Vary kernel sparsity weight from λ = 0.0 to λ = 5.0 (ablation) | BlindDPS · FFHQ Gaussian deblurring | MNC: 0.996 -> 0.997 | No clear effect |
| Vary kernel sparsity weight from λ = 0.0 to λ = 5.0 (ablation) | BlindDPS · FFHQ Gaussian deblurring | MSE: 0.000 -> 0.000 | No clear effect |
Closest papers by meaning
- A Diffusion Model with State Estimation for Degradation-Blind Inverse Imaging (AAAI 2024 · similarity 0.90)
- Blind Image Restoration via Fast Diffusion Inversion (NeurIPS 2024 · similarity 0.90)
- A Variational Perspective on Solving Inverse Problems with Diffusion Models (ICLR 2024 · similarity 0.89)
- Unleashing the Denoising Capability of Diffusion Prior for Solving Inverse Problems (NeurIPS 2024 · similarity 0.89)
- Integrating Intermediate Layer Optimization and Projected Gradient Descent for Solving Inverse Problems with Diffusion Models (ICML 2025 · similarity 0.89)
- Improving Diffusion-based Inverse Algorithms under Few-Step Constraint via Linear Extrapolation (NeurIPS 2025 · similarity 0.89)
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.