Research map / Image generation / Image restoration
Towards Authentic Face Restoration with Iterative Diffusion Models and Beyond
Yang Zhao, Tingbo Hou, Yu-Chuan Su, Xuhui Jia, Yandong Li, Matthias Grundmann
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 |
|---|---|---|---|
| apply extrinsic iteration: restore the training data and retrain | FFHQ dataset · FFHQ | NIMA score: 4.51 -> 4.85 | Improved |
| increase diffusion sampling steps beyond 10 (ablation) | IDM · CelebA-HQ test (original) | PSNR: 27.42 -> 27.50 | No clear effect |
| use 10 instead of 5 sampling steps (ablation) | IDM · CelebA-HQ test (original) | PSNR: 13.11 -> 27.42 | Improved |
| use DDIM sampling instead of DDPM (ablation) | IDM · face restoration | restoration quality | Hurt |
| predict x0 directly instead of noise in first iteration (ablation) | IDM · training | convergence speed: approximately 90% faster | Improved |
| use L1 loss instead of L2 loss (ablation) | IDM · FFHQ | color faithfulness | Improved |
| train StyleGAN2 on IDM-restored FFHQ | StyleGAN2 · FFHQ 512x512 | FID: 3.08 -> 2.78 | Improved |
| train BigGAN on IDM-restored ImageNet | BigGAN · ImageNet 128x128 | FID: 9.77 -> 8.97 | Improved |
| train diffusion model on IDM-restored ImageNet | conditional DDM · ImageNet 128x128 (100K iterations) | FID: 14.43 -> 11.10 | Improved |
| extrinsic iteration (retrain on restored data) (ablation) | IDM · restoration quality | user preference | Improved |
Closest papers by meaning
- ReF-LDM: A Latent Diffusion Model for Reference-based Face Image Restoration (NeurIPS 2024 · similarity 0.91)
- OSDFace: One-Step Diffusion Model for Face Restoration (CVPR 2025 · similarity 0.90)
- FaceMe: Robust Blind Face Restoration with Personal Identification (AAAI 2025 · similarity 0.90)
- Self-Supervised Selective-Guided Diffusion Model for Old-Photo Face Restoration (NeurIPS 2025 · similarity 0.89)
- Overcoming False Illusions in Real-World Face Restoration with Multi-Modal Guided Diffusion Model (ICLR 2025 · similarity 0.89)
- $InterLCM$: Low-Quality Images as Intermediate States of Latent Consistency Models for Effective Blind Face Restoration (ICLR 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.
