Research map / Image generation / Image restoration
Degradation-Modeled Multipath Diffusion for Tunable Metalens Photography
Jianing Zhang, Jiayi Zhu, Feiyu Ji, Xiaokang Yang, Xiaoyun Yuan
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 |
|---|---|---|---|
| raise diffusion injection factor alpha from 0 to 1.05 (ablation) | DMDiff · Metalens test set | MUSIQ: 31.24 -> 51.85 | Improved |
| raise diffusion injection factor alpha from 0 to 1.05 (ablation) | DMDiff · Metalens test set | LPIPS: 0.2715 -> 0.1485 | Improved |
| raise diffusion injection factor alpha from 0.5 to 1.05 (ablation) | DMDiff · Metalens test set | PSNR: 30.39 -> 29.75 | Hurt |
| raise diffusion injection factor alpha from 0.5 to 1.05 (ablation) | DMDiff · Metalens test set | SSIM: 0.8743 -> 0.8598 | Hurt |
| add FWHM degradation cue (SVDA) to the model (ablation) | DMDiff · Metalens test set | PSNR: 26.62 -> 29.89 | Improved |
| add FWHM degradation cue (SVDA) to the model (ablation) | DMDiff · Metalens test set | MUSIQ: 50.55 -> 50.63 | No clear effect |
| add negative prompt path (ablation) | DMDiff · Metalens test set | MUSIQ: 44.15 -> 50.63 | Improved |
| add negative prompt path (ablation) | DMDiff · Metalens test set | PSNR: 28.21 -> 29.89 | Improved |
| add all proposed modules to plain LoRA fine-tuned base (ablation) | SD-Turbo · Metalens test set | PSNR: 17.12 -> 29.89 | Improved |
| use DMDiff instead of SwinIR | metalens restoration · Metalens test set | MUSIQ: 36.86 -> 51.85 | Improved |
| use DMDiff instead of SeeSR-s50 | metalens restoration · Metalens test set | PSNR: 23.95 -> 30.31 | Improved |
| use DMDiff instead of SwinIR | metalens restoration · Metalens test set | SSIM: 0.8786 -> 0.8731 | Hurt |
Closest papers by meaning
- Diffusion Once and Done: Degradation-Aware LoRA for All-in-One Image Restoration (AAAI 2026 · similarity 0.86)
- Towards Illumination-Aware Restoration of Metalens-Captured Images: A New Dataset and a Strong Baseline (AAAI 2026 · similarity 0.86)
- Unsupervised Diffusion-Based Degradation Modeling for Real-World Super-Resolution (AAAI 2025 · similarity 0.85)
- AGLLDiff: Guiding Diffusion Models Towards Unsupervised Training-free Real-world Low-light Image Enhancement (AAAI 2025 · similarity 0.84)
- GenDeg: Diffusion-based Degradation Synthesis for Generalizable All-In-One Image Restoration (CVPR 2025 · similarity 0.84)
- FAM Diffusion: Frequency and Attention Modulation for High-Resolution Image Generation with Stable Diffusion (CVPR 2025 · similarity 0.84)
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.
