mlsanityData to 14 Oct 2026

FlowSGS: Improving Flow Matching Priors for Inverse Imaging with Stochastic Interpolants

arXiv 2609.20769 ↗First posted 17 Sep 2026cs.CVcited by 0

On the map: Diffusion models → Inverse problems and image restoration also fits: Flow matching, time series and science

Abstract (arXiv)

Flow matching has emerged as the state-of-the-art generative model and has been used for plug-and-play (PnP) priors to solve inverse problems in computational imaging. However, existing flow-based inverse solvers make simplifying approximations in posterior sampling and/or assume linear forward models. In contrast, principled diffusion-based methods exist but are computationally expensive. To circumvent these problems, we introduce FlowSGS, a flow-based posterior sampling method using Split Gibbs Sampling (SGS) with annealed coupling, and make flow-specific interventions to the prior sampling and annealing steps, theoretically grounded in the Stochastic Interpolants (SI) framework. Using a novel timestep correction and accelerated warm-up with one-step flow models, our method is faster and more accurate than the state of the art across a range of imaging problems. We demonstrate the first use of flow-based samplers for a nonlinear inverse problem (Fourier phase retrieval) and apply one-step flow models to principled Bayesian inverse imaging.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in Inverse problems and image restoration, one of 14 directions in Diffusion models. Builds on 54 of its 71 references that are in the map. First posted on arXiv 17 Sep 2026.

Builds on

54 of 71 references.

Closest ancestors · 12

Same direction, newest first

  1. 2025PRISM: Probabilistic and Robust Inverse Solver with Measurement-Conditioned Diffusion Prior for Blind Inverse Problems19 Sep
  2. 2025InverseBench: Benchmarking Plug-and-Play Diffusion Priors for Inverse Problems in Physical Sciences14 Mar
  3. 2025Split Gibbs Discrete Diffusion Posterior Sampling3 Mar
  4. 2024PnP-Flow: Plug-and-Play Image Restoration with Flow Matching3 Oct
  5. 2024Improving Diffusion Inverse Problem Solving with Decoupled Noise Annealing1 Jul
  6. 2024Principled Probabilistic Imaging using Diffusion Models as Plug-and-Play Priors29 May
  7. 2024Provably Robust Score-Based Diffusion Posterior Sampling for Plug-and-Play Image Reconstruction25 Mar
  8. 2023Denoising Diffusion Models for Plug-and-Play Image Restoration15 May
  9. 2023A Variational Perspective on Solving Inverse Problems with Diffusion Models7 May
  10. 2022Diffusion Posterior Sampling for General Noisy Inverse Problems29 Sep
  11. 2022Improving Diffusion Models for Inverse Problems using Manifold Constraints2 Jun
  12. 2022Denoising Diffusion Restoration Models27 Jan

Other references · 42

  1. 2020Denoising Diffusion Probabilistic Modelscited by 21,076
  2. 2020Score-Based Generative Modeling through Stochastic Differential Equationscited by 9,122
  3. 2018A Style-Based Generator Architecture for Generative Adversarial Networkscited by 6,013
  4. 2022Flow Matching for Generative Modelingcited by 5,702
  5. 2024Scaling Rectified Flow Transformers for High-Resolution Image Synthesiscited by 4,181
  6. 2019Generative Modeling by Estimating Gradients of the Data Distributioncited by 4,101
  7. 2022Flow Straight and Fast: Learning to Generate and Transfer Data with Rectified Flowcited by 3,700
  8. 2014Deep Learning Face Attributes in the Wildcited by 3,693
  9. 2022Elucidating the Design Space of Diffusion-Based Generative Modelscited by 3,045
  10. 2021Convex optimizationcited by 2,515
  11. 2019StarGAN v2: Diverse Image Synthesis for Multiple Domainscited by 1,098
  12. 2022Building Normalizing Flows with Stochastic Interpolantscited by 995
30 more
  1. 2023Stochastic Interpolants: A Unifying Framework for Flows and Diffusionscited by 939
  2. 2024SiT: Exploring Flow and Diffusion-based Generative Models with Scalable Interpolant Transformerscited by 737
  3. 2025Mean Flows for One-step Generative Modelingcited by 561
  4. 2021Solving Inverse Problems in Medical Imaging with Score-Based Generative Modelscited by 521
  5. 2018fastMRI: An Open Dataset and Benchmarks for Accelerated MRIcited by 470
  6. 2020Plug-and-Play Image Restoration with Deep Denoiser Priorcited by 470
  7. 2019Algorithm Unrolling: Interpretable, Efficient Deep Learning for Signal and Image Processingcited by 421
  8. 2021Score-based diffusion models for accelerated MRIcited by 311
  9. 2021Robust Compressed Sensing MRI with Deep Generative Priorscited by 286
  10. 2016Plug-and-Play ADMM for Image Restoration: Fixed Point Convergence and Applicationscited by 246
  11. 2022Diffusion models as plug-and-play priorscited by 231
  12. 2019Rapid Convergence of the Unadjusted Langevin Algorithm: Isoperimetry Sufficescited by 227
  13. 2024D-Flow: Differentiating through Flows for Controlled Generationcited by 122
  14. 2022Plug-and-Play Methods for Integrating Physical and Learned Models in Computational Imagingcited by 96
  15. 2023Minimizing Trajectory Curvature of ODE-based Generative Modelscited by 89
  16. 2023Training-free Linear Image Inverses via Flowscited by 72
  17. 2025FlowDPS: Flow-Driven Posterior Sampling for Inverse Problemscited by 68
  18. 2020Deep Probabilistic Imaging: Uncertainty Quantification and Multi-modal Solution Characterization for Computational Imagingcited by 57
  19. 2024Flow Priors for Linear Inverse Problems via Iterative Corrupted Trajectory Matchingcited by 53
  20. 2018Split-and-augmented Gibbs sampler - Application to large-scale inference problemscited by 47
  21. 2023Plug-and-Play split Gibbs sampler: embedding deep generative priors in Bayesian inferencecited by 42
  22. 2026Meta Flow Maps enable scalable reward alignmentcited by 37
  23. 2023Generative Plug and Play: Posterior Sampling for Inverse Problemscited by 24
  24. 2025Solving Inverse Problems with FLAIRcited by 17
  25. 2015Computational Imaging for VLBI Image Reconstructioncited by 12
  26. 2022The split Gibbs sampler revisited: improvements to its algorithmic structure and augmented target distributioncited by 11
  27. 2024Regularization by denoising: Bayesian model and Langevin-within-split Gibbs samplingcited by 9
  28. 2024Event-horizon-scale Imaging of M87* under Different Assumptions via Deep Generative Image Priorscited by 7
  29. 2025Latent Refinement via Flow Matching for Training-free Linear Inverse Problem Solvingcited by 2
  30. 2026ShuffleFlow: Scalable Posterior Inference for Bayesian Inverse Imagingcited by 1

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026What Do Flow-Based Inverse Solvers Approximate? A Posterior-Transport View
  2. 2026Fast and Faithful: Principled Conditional Flow Matching for Inverse Problems
  3. 2026Flow Matching for Fast Posterior Sampling in Bayesian Inverse Problems
  4. 2024DAWN-FM: Data-Aware and Noise-Informed Flow Matching for Solving Inverse Problems
  5. 2026P-Flow: Proxy-gradient Flows for Linear Inverse Problems
  6. 2024Fast Samplers for Inverse Problems in Iterative Refinement Models
  7. 2026D-Flow SGLD: Source-Space Posterior Sampling for Scientific Inverse Problems with Flow Matching
  8. 2026Kinetic Langevin Meets Split Gibbs: Accelerated Posterior Sampling for Imaging Inverse Problems with Diffusion Priors