Research map / Image generation / Fast diffusion sampling
FlowDPS : Flow-Driven Posterior Sampling for Inverse Problems
Jeongsol Kim, Bryan Sangwoo Kim, Jong Chul Ye
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 |
|---|---|---|---|
| add stochastic noise to the noise estimate (ablation) | FlowDPS · AFHQ SR x12 (Avgpool) | FID: 32.85 -> 16.85 | Improved |
| add stochastic noise to the noise estimate (ablation) | FlowDPS · FFHQ SR x12 (Avgpool) | FID: 78.67 -> 33.78 | Improved |
| add stochastic noise to the noise estimate (ablation) | FlowDPS · AFHQ SR x12 (Avgpool) | PSNR: 24.75 -> 25.88 | Improved |
| add stochastic noise to the noise estimate (ablation) | FlowDPS · FFHQ SR x12 (Bicubic) | FID: 82.07 -> 33.75 | Improved |
| add stochastic noise to the noise estimate (ablation) | FlowDPS · AFHQ SR x12 (Bicubic) | LPIPS: 0.268 -> 0.188 | Improved |
| use interpolation gamma=sigma_t instead of gamma=1 (full data-consistency update) (ablation) | FlowDPS · Linear inverse problems (qualitative) | visual sharpness | Improved |
| use gamma=sigma_t instead of gamma=1-sigma_t (ablation) | FlowDPS · Linear inverse problems (qualitative) | trajectory consistency | Improved |
| use FlowDPS instead of FlowChef | FlowDPS · DIV2K 768x768, SR x12 (Avgpool) | FID: 101.5 -> 58.29 | Improved |
| use FlowDPS instead of ReSample | FlowDPS · FFHQ 768x768, SR x12 (Avgpool) | FID: 102.7 -> 33.78 | Improved |
| raise CFG scale at higher NFE (28, 50, 100 with lambda 0.0, 2.0, 7.5) (ablation) | FlowDPS · DIV2K (qualitative) | high-frequency detail | Mixed |
Later papers that compared with this method
8 later accepted papers report FlowDPS as a baseline in their results tables; 7 of them report a better number on most of the results they share with it.
- Triadic Dynamics Aware Diffusion Posterior Sampling for Inverse Problems: Optimizing Guidance and Stochasticity Schedules (ICML 2026 · better on 4 of 4 shared results)
- Variational Flow Maps: Make Some Noise for One-Step Conditional Generation (ICML 2026 · better on 2 of 4 shared results)
- Trajectory Stitching for Solving Inverse Problems with Flow-Based Models (ICML 2026 · better on 3 of 3 shared results)
- SONIC: Spectral Optimization of Noise for Inpainting with Consistency (ECCV 2026 · better on 3 of 3 shared results)
- Stochastic Optimal Control Sampling for Diffusion Inverse Problems (ECCV 2026 · better on 2 of 2 shared results)
- Image Restoration via Diffusion Models with Dynamic Resolution (ICML 2026 · better on 1 of 1 shared result)
- DICT: Data Injection and Contrastive Trajectory Refinement for Conditional Image Generation with Diffusion Models (ECCV 2026 · better on 1 of 1 shared result)
- Solving Inverse Problems with FLAIR (NeurIPS 2025 · better on 3 of 3 shared results)
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- FlowChef — FlowChef: Steering of Rectified Flow Models for Controlled Generations (ICCV 2025) · better on 9 of 9 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Latent Refinement via Flow Matching for Training-free Linear Inverse Problem Solving (NeurIPS 2025 · similarity 0.93)
- Solving Inverse Problems with FLAIR (NeurIPS 2025 · similarity 0.92)
- Flow Priors for Linear Inverse Problems via Iterative Corrupted Trajectory Matching (NeurIPS 2024 · similarity 0.92)
- Fast samplers for Inverse Problems in Iterative Refinement models (NeurIPS 2024 · similarity 0.92)
- PnP-Flow: Plug-and-Play Image Restoration with Flow Matching (ICLR 2025 · similarity 0.91)
- Inverse Flow and Consistency Models (ICML 2025 · similarity 0.90)
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.
