Research map / Image generation / Image restoration
Stage-wise Distortion–Perception Traversal in Zero-shot Inverse Problems with Diffusion Models
Jiawei Zhang, Ziyuan Liu, Leon Yan, Zhenyu Xiao, Yuantao Gu
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 re-noising time t0 from 0 to 600 (ablation) | LMAP-RPS · MS-COCO inpainting | PSNR: 28.14 -> 27.22 | Hurt |
| raise re-noising time t0 from 0 to 600 (ablation) | LMAP-RPS · MS-COCO inpainting | FID: 61.35 -> 92.23 | Hurt |
| raise re-noising time t0 from 0 to 600 (ablation) | LMAP-RPS · MS-COCO SR 4x | FID: 107.57 -> 87.20 | Improved |
| raise re-noising time t0 from 0 to 600 (ablation) | LMAP-RPS · MS-COCO SR 4x | PSNR: 25.03 -> 24.49 | Hurt |
| raise re-noising time t0 from 0 to 600 (ablation) | LMAP-RPS · MS-COCO HDR | PSNR: 25.86 -> 23.06 | Hurt |
| raise re-noising time t0 from 0 to 600 (ablation) | LMAP-RPS · MS-COCO HDR | FID: 108.79 -> 116.08 | Hurt |
| raise re-noising time t0 from 0 to 600 (ablation) | LMAP-RPS · MS-COCO CS 2x | FID: 113.06 -> 113.58 | No clear effect |
| raise re-noising time t0 from 0 to 600 (ablation) | LMAP-RPS · MS-COCO nonlinear deblur | FID: 119.43 -> 118.69 | No clear effect |
| raise re-noising time t0 from 0 to 600 (ablation) | LMAP-RPS · MS-COCO anisotropic deblur | FID: 90.80 -> 85.01 | Improved |
| use LMAP-RPS (t0=0) instead of Latent-DPS | Latent-DPS · MS-COCO inpainting | inference time (s/image): 242 -> 45 | Improved |
| use LMAP-RPS (t0=0) instead of Latent-DPS | Latent-DPS · MS-COCO SR 4x | inference time (s/image): 250 -> 23 | Improved |
| use LMAP-RPS (t0=0) instead of Latent-SITCOM | Latent-SITCOM · MS-COCO inpainting | PSNR: 28.06 -> 28.14 | Improved |
Closest papers by meaning
- DiracDiffusion: Denoising and Incremental Reconstruction with Assured Data-Consistency (ICML 2024 · similarity 0.90)
- A Variational Perspective on Solving Inverse Problems with Diffusion Models (ICLR 2024 · similarity 0.89)
- SITCOM: Step-wise Triple-Consistent Diffusion Sampling For Inverse Problems (ICML 2025 · similarity 0.89)
- InverseBench: Benchmarking Plug-and-Play Diffusion Priors for Inverse Problems in Physical Sciences (ICLR 2025 · similarity 0.89)
- Analyzing and Guiding Zero-Shot Posterior Sampling in Diffusion Models (NeurIPS 2026 · similarity 0.88)
- Traversing Distortion-Perception Tradeoff using a Single Score-Based Generative Model (CVPR 2025 · 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.
