PnP-CF: A Consensus Flow-Matching Solver for General Inverse Problems

Dibyanshu Kumar; Magda Gregorova

cs.LGOct 19, 2026

PnP-CF: A Consensus Flow-Matching Solver for General Inverse Problems

Authors: Dibyanshu Kumar, Magda Gregorova

Organizations: TH Würzburg-Schweinfurt, CAIRO, University of Würzburg

Abstract

Plug-and-Play (PnP) methods using continuous-time deep generative models as learned data distribution priors enable synthesizing high-quality solutions of model-based inverse problems. The quality of the solutions depends critically on the interplay and consensus between the inverse problem solver and the generative model sampler. We revisit a recently introduced flow-based PnP framework, which we refer to as PnP-CF (Consensus Flow), and extend its application to nonlinear inverse problems, including those with learned neural-network forward operators. PnP-CF combines a flow-matching denoiser with (i) ADMM-like dual variable updates and (ii) a time-adaptive pullback onto the intermediate flow probability path that combines a stochastic and a deterministic component. We discuss how these two features stabilize the algorithm and promote consensus, and demonstrate on a set of experiments that PnP-CF not only solves standard linear inverse problems with quality at least comparable to state-of-the-art baselines, but also extends gracefully to nonlinear inverse problems, including forward operators realized by pretrained neural networks, for which existing methods produce unsatisfactory results or fail completely.

Figures & tables

Figure 1: Phases of PnP-CF. Left: Starting from $\tilde{\mathbf{z}}$ , the denoiser $\mathcal{D}(\tilde{\mathbf{z}})$ produces $\hat{\mathbf{z}}$ (green). The dual correction (gray) shifts $\hat{\mathbf{z}}$ by ${\mathbf{u}}$ , and the forward-consistency operator $\mathcal{F}({\mathbf{y}},\tilde{\mathbf{x}})$ maps it to $\hat{\mathbf{x}}$ (blue), pulling it towards the region consistent with ${\mathbf{y}}$ under $l({\mathbf{y}};{\mathbf{x}})$ . Right: The time-adaptive pullback constructs the next $\tilde{\mathbf{z}}$ . The deterministic component ${\mathbf{\epsilon}}_{d}$ , computed from the previous $\tilde{\mathbf{z}}$ and $\hat{\mathbf{z}}$ , combines with ${\mathbf{\epsilon}}_{s}\sim p_{t=0}$ to form ${\mathbf{\epsilon}}$ . The dual-corrected point $\hat{\mathbf{x}}+{\mathbf{u}}$ (gray) and the ${\mathbf{\epsilon}}$ together produces the new $\tilde{\mathbf{z}}$ , which serves as the input to the denoiser at the next iteration.

Panel of Figure 2 — Figure 2: Ablation of PnP-CF update strategies on the box-inpainting inverse problem. Left: qualitative comparison of stochastic-only, deterministic-only, and time-adaptive pullback strategies. Right: effect of the dual variable ${\mathbf{u}}$ on the forward-consistency and denoising updates, shown through PSNR and SSIM trajectories over iterations, averaged over 100 images.

Figure 3: Evolution of the primal residual $\|\hat{\mathbf{x}}^{k}-\hat{\mathbf{z}}^{k}\|_{2}$ across linear and nonlinear inverse problems on AFHQ-Cat (left) and CelebA (right).

Figure 4: Comparison of image restoration methods on box inpainting, super-resolution, and random inpainting, with additional qualitative results provided in appendix F . Our method yields visually sharp reconstructions while preserving structural consistency across tasks.

Method	Time (s)	Mem (GB)
PnP-CF (ours)	17.62	0.30
PnP-Flow	16.40	0.30
D-Flow	71.42	13.34
FlowDPS	11.14	0.32

Table 2: Per batch runtime and peak GPU memory (HDR recon. - CelebA)

Panel of Figure 5 — Figure 5: Qualitative results. Left: Segmentation-guided reconstruction results for D-Flow and PnP-CF. Right: Qualitative reconstruction on nonlinear inverse problems, including HDR reconstruction, phase retrieval, and nonlinear deblurring with additional results in appendix G .

Method	LPIPS $\downarrow$	SSIM $\uparrow$	mIoU $\uparrow$
PnP-CF	0.27	0.41	0.99
PnP-Flow	0.35	0.28	0.37
D-Flow	0.38	0.11	0.98
FlowDPS	0.36	0.28	0.33

Table 3: Segmentation-guided reconstruction results.

Appendix figures & tables14 assets

Supplementary material from the paper’s appendix.

Appendix

Dataset	Task	$K$	$\rho$	$\beta$	$\gamma$	Denoiser calls	GD steps
CelebA	Box inpainting	100	50	2.5	1.0	5	–
CelebA	Random inpainting	100	100	2.5	1.0	5	–
CelebA	Super-resolution	100	50	1.3	1.0	5	–
CelebA	Phase retrieval	100	1500	1.8	1.5	5	10
CelebA	HDR reconstruction	100	100	1.5	0.2	5	10
CelebA	Nonlinear deblurring	200	100	1.5	1.5	1	5

Appendix

Table 4: Hyperparameter settings per task and dataset for the time-adaptive PnP-CF variant.

K

: number of PnP-CF iterations ( steps_pnp );

\rho

: base ADMM penalty ( rho_admm );

\beta

: penalty-schedule exponent ( beta_rho );

\gamma

: time-schedule exponent ( t_schedule_gamma ); Denoiser calls: number of denoiser samples averaged per iteration ( num_denoiser_calls ); GD steps: nonlinear x-update gradient steps ( admm_gd_steps , linear tasks use CG instead).

	Box Inpaint.	Super-res.	Rand. Inpaint.	HDR Recon.	Phase Retr.	Nonl. Deblur.
OT-ODE
$t_{0}$ (start time)	0.1	0.1	0.1	N/A	N/A	N/A
$\gamma$ schedule	$\gamma_{t}$	constant	constant	N/A	N/A	N/A
$K$ (steps)	100	100	100	N/A	N/A	N/A
Flow Priors
$\lambda$ (regularization)	10,000	10,000	10,000	N/A	N/A	N/A

Appendix

Table 5: Hyperparameters used for all baseline methods on the CelebA dataset.

	Box Inpaint.	Super-res.	Rand. Inpaint.	HDR Recon.	Phase Retr.	Nonl. Deblur.
OT-ODE
$t_{0}$ (start time)	0.1	0.1	0.1	N/A	N/A	N/A
$\gamma$ schedule	$\gamma_{t}$	constant	constant	N/A	N/A	N/A
$K$ (steps)	100	500	100	N/A	N/A	N/A
Flow Priors
$\lambda$ (regularization)	10,000	10,000	10,000	N/A	N/A	N/A

Appendix

Table 6: Hyperparameters used for all baseline methods on the AFHQ-Cat dataset.

Figure 6: Effect of the dual-variable update on reconstruction quality across optimization iterations. PSNR (top), SSIM (middle), and LPIPS (bottom) are shown for super-resolution, random inpainting, phase retrieval, and HDR reconstruction on CelebA, comparing PnP-CF ‘with’ and ‘without’ the dual-variable update. Across tasks, the dual-variable formulation generally improves convergence speed and final reconstruction quality, with particularly large gains for super-resolution, random inpainting, and HDR reconstruction.

		CelebA			AFHQ-Cat
Task	Method	PSNR $\uparrow$	LPIPS $\downarrow$	SSIM $\uparrow$	PSNR $\uparrow$	LPIPS $\downarrow$	SSIM $\uparrow$
HDR Reconstruction	PnP-CF (adaptive)	27.59	0.04	0.85	26.29	0.08	0.86
	PnP-CF (stochastic)	22.44	0.10	0.80	22.97	0.32	0.71
	PnP-CF (deterministic)	18.12	0.36	0.43	24.11	0.13	0.75
Box Inpainting	PnP-CF (adaptive)	26.36	0.05	0.89	22.65	0.13	0.82
	PnP-CF (stochastic)	22.79	0.18	0.74	21.24	0.37	0.68

Appendix

Table 8: Task-wise performance of the three PnP-CF pullback variants (time-adaptive, stochastic, deterministic) across all six inverse tasks. Best result per task and metric, among the three variants, is shown in bold .

Panel of Figure 7 — Figure 7: Additional qualitative ablation results comparing the pullback variants of PnP-CF on CelebA (left) and AFHQ-Cat (right). Within each dataset, columns from left to right show the clean image, degraded observation, stochastic-only ( ${\mathbf{\epsilon}}_{s}$ ), deterministic-only ( ${\mathbf{\epsilon}}_{d}$ ), and time-adaptive ( ${\mathbf{\epsilon}}_{s},{\mathbf{\epsilon}}_{d}$ ) variants. From top to bottom, the rows correspond to box inpainting, random inpainting, super-resolution, phase retrieval, nonlinear deblurring, and HDR reconstruction.

Figure 8: Failure cases for segmentation-guided reconstruction. Left: reference images. Middle: PnP-CF reconstructions. Right: reconstructed images with the predicted segmentation boundaries overlaid in red. Despite achieving strong segmentation consistency, some reconstructions exhibit distorted facial geometry and perceptual artifacts. These examples illustrate that satisfying a learned segmentation forward model does not necessarily guarantee a visually plausible reconstruction.

		PSNR $\uparrow$		SSIM $\uparrow$		LPIPS $\downarrow$
Dataset	Task	Rec	Noisy	Rec	Noisy	Rec	Noisy
CelebA	Box Inpainting	24.71	18.91	0.76	0.66	0.08	0.31
	Random Inpainting	30.15	11.46	0.90	0.18	0.04	1.09
	Super Resolution	20.28	10.71	0.56	0.18	0.24	0.82
	HDR Reconstruction	21.97	17.03	0.79	0.50	0.06	0.11
	Nonlinear Deblurring	25.36	24.77	0.73	0.66	0.15	0.24

Appendix

Table 9: PnP-CF results on reconstructed vs. noisy/degraded inputs with default (non-scheduled) hyperparameter settings on CelebA and AFHQ-Cat.

		Box inpainting			Super-resolution			Random inpainting
Dataset	Method	PSNR $\uparrow$	LPIPS $\downarrow$	SSIM $\uparrow$	PSNR $\uparrow$	LPIPS $\downarrow$	SSIM $\uparrow$	PSNR $\uparrow$	LPIPS $\downarrow$	SSIM $\uparrow$
CelebA	PnP-CF (ours)	31.71	0.02	0.95	31.37	0.03	0.86	33.99	0.02	0.95
	PnP-Flow	30.82	0.04	0.94	31.31	0.06	0.91	33.92	0.02	0.95
	FLOWER	31.66	0.02	0.95	32.88	0.04	0.90	33.91	0.03	0.95
	FlowDPS	28.48	0.06	0.89	25.94	0.12	0.79	26.42	0.11	0.80
	D-Flow	28.28	0.06	0.85	29.40	0.04	0.83	31.96	0.03	0.91

Appendix

Table 10: Reconstruction quality on CelebA and AFHQ-Cat for linear inverse problems under the milder degradation settings used in Martin et al. [2025] . Best results are shown in bold .

Figure 9: Additional qualitative comparisons on the CelebA dataset across linear inverse tasks, including box inpainting (1st row), random inpainting (2nd row) and super-resolution (3rd row).

Figure 10: Additional qualitative comparisons on the AFHQ-Cat dataset across linear inverse tasks, including box inpainting (1st row), random inpainting (2nd row) and super-resolution (3rd row).

Figure 11: Additional qualitative comparisons on the CelebA dataset across nonlinear inverse tasks, including phase retrieval (1st row), nonlinear deblurring (2nd row) and HDR reconstruction (3rd row).

Figure 12: Additional qualitative comparisons on the AFHQ-Cat dataset across nonlinear inverse tasks, including phase retrieval (1st row), nonlinear deblurring (2nd row) and HDR reconstruction (3rd row).

Figure 13: Qualitative comparison for segmentation-guided reconstruction on CelebA-HQ. From left to right: reference image, target segmentation mask, and reconstructions from PnP-Flow, FlowDPS, D-Flow, and PnP-CF. PnP-Flow exhibits mode collapse across different masks, while FlowDPS produces plausible but weakly conditioned reconstructions. D-Flow follows the segmentation constraint more closely but introduces severe visual artifacts. In contrast, PnP-CF produces visually plausible reconstructions while maintaining strong consistency with the target segmentation.

Inverse Problems

Inverse problems infer hidden causes, parameters, or signals from indirect, noisy measurements using physical models, priors, and statistical methods.

See all 246 papers

FlowADMM: Plug-and-play ADMM with Flow-based Renoise-Denoise Priors
May 9, 2026Hendrik Sommerhoff, Michael MoellerAlternating Direction Method of Multipliers Image Restoration
Computer Vision Group, University of Siegen Hölderlinstraße 3, 57076 Siegen, Germany
Stochastic Generative Plug-and-Play Priors
Apr 4, 2026Chicago Y. Park, Edward P. Chandler, Yuyang Hu +4Image Inverse Problems Diffusion-Based Inverse Problems
University of Wisconsin–Madison · Washington University in St. Louis · Los Alamos National Laboratory
FlowSGS: Improving Flow Priors for Inverse Imaging with Stochastic Interpolants
Sep 17, 2026Tianao Li, Xinhui Qian, Emma AlexanderFlow Matching Markov Chain Monte Carlo
Department of Computer Science, Northwestern University · NSF-Simons AI Institute for the Sky (SkAI) · Department of Statistics and Data Science, Northwestern University

May 9, 2026J/K move · Enter open · S save

FlowADMM: Plug-and-play ADMM with Flow-based Renoise-Denoise Priors

Hendrik Sommerhoff, Michael Moeller

Computer Vision Group, University of Siegen Hölderlinstraße 3, 57076 Siegen, Germany

Alternating Direction Method of Multipliers Image Restoration Constrained Optimization Inverse Problems

Open paper AskPDF arXiv

Plug-and-play (PnP) methods for solving inverse problems have recently achieved strong performance by leveraging denoising priors based on powerful generative diffusion and flow models. However, existing diffusion- and flow-based PnP methods typically rely on stochastic renoise-denoise operations, which complicate the analysis of their convergence behavior. In this work, we identify and formalize the deterministic renoise-denoise operator underlying flow-based plug-and-play methods. This perspective reveals that these methods implicitly define a deterministic operator given by the expectation of a denoiser over the latent noise distribution. Building on this insight, we propose FlowADMM, a PnP algorithm that integrates the renoise-denoise operator into the classical alternating direction method of multiplier (ADMM) framework. We establish convergence guarantees for FlowADMM under weak Lipschitz conditions on the underlying flow network, and extend the analysis to non-stationary time schedules. Empirically, FlowADMM achieves state-of-the-art performance among flow-based PnP methods on a range of inverse problems, including denoising, deblurring, super-resolution, and inpainting, while requiring fewer data consistency evaluations than prior approaches.

Abstract

Inverse Problems

Explore similar work

FlowADMM: Plug-and-play ADMM with Flow-based Renoise-Denoise Priors

Stochastic Generative Plug-and-Play Priors

FlowSGS: Improving Flow Priors for Inverse Imaging with Stochastic Interpolants