Research map / Image generation / Fast diffusion sampling
SLAIR: Structured Latent Flow Matching for All-in-One Image Restoration
Shuyi Liang, Yixin Yang, Hanyue Lou, Yuning Cui, Boxin Shi
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 |
|---|---|---|---|
| finetune the VAE with a classification-based latent separation loss (ablation) | SLAIR VAE · ImageNet VAE reconstruction | rFID: 0.278 -> 0.225 | Improved |
| finetune the VAE with a classification-based latent separation loss (ablation) | SLAIR VAE · ImageNet VAE reconstruction | PSNR: 25.79 -> 25.99 | Improved |
| latent flow matching restoration (SLAIR) versus best baseline | All-in-one 8-task · All-in-one 8-task (setting i, average of 9 datasets) | PSNR: 24.77 -> 26.15 | Improved |
| SLAIR versus baselines on single weather benchmarks | Rain1400, Raindrop, UHDSnow · Rain1400 (setting i) | PSNR: 30.19 (DiffUIR) vs 29.36 (SLAIR) | Mixed |
| SLAIR versus MoCE-IR in the standard 5-task setting | Standard 5-task · Standard 5-task all-in-one (setting ii, average) | PSNR: 28.52 -> 28.39 PSNR; 52.18 -> 31.05 FID | Mixed |
| add wavelet high-frequency refinement loss in decoder (ablation) | SLAIR decoder · qualitative (Fig. 4) | visual residual | Improved |
Closest papers by meaning
- Reversing Flow for Image Restoration (CVPR 2025 · similarity 0.89)
- Latent Refinement via Flow Matching for Training-free Linear Inverse Problem Solving (NeurIPS 2025 · similarity 0.87)
- Cross Flow: One-Step Generation Across Latent and Pixel Spaces (NeurIPS 2026 · similarity 0.87)
- Posterior-Mean Rectified Flow: Towards Minimum MSE Photo-Realistic Image Restoration (ICLR 2025 · similarity 0.86)
- PnP-Flow: Plug-and-Play Image Restoration with Flow Matching (ICLR 2025 · similarity 0.86)
- Residual-Conditioned Optimal Transport: Towards Structure-Preserving Unpaired and Paired Image Restoration (ICML 2024 · similarity 0.86)
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.
