Product-of-Gaussian-Mixture Diffusion Models for Joint Nonlinear MRI Reconstruction

Laurenz Nagler; Martin Zach; Thomas Pock

arxiv: 2605.10629 · v1 · submitted 2026-05-11 · 💻 cs.CV

Product-of-Gaussian-Mixture Diffusion Models for Joint Nonlinear MRI Reconstruction

Laurenz Nagler , Martin Zach , Thomas Pock This is my paper

Pith reviewed 2026-05-12 05:12 UTC · model grok-4.3

classification 💻 cs.CV

keywords MRI reconstructiondiffusion modelscoil sensitivity estimationjoint reconstructionGaussian mixtureimage priorsnonlinear optimization

0 comments

The pith

A compact product-of-Gaussian-mixture diffusion model acts as an image prior for joint reconstruction of MRI images and coil sensitivities.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper develops a reconstruction technique that estimates both the MRI image and the coil sensitivity maps in one optimization step. It replaces large time-conditioned networks with a parameter-efficient diffusion model that serves as the image prior and pairs it with a standard smoothness constraint on the coil maps. This setup removes the usual requirement for offline coil calibration and makes the method less sensitive to changes in image contrast, body region, or k-space trajectory. A refined parameterization of the diffusion prior further raises performance on denoising and full reconstruction tasks.

Core claim

The authors show that a product-of-Gaussian-mixture diffusion model can serve as an effective, low-parameter image prior when jointly optimizing the image and coil sensitivities under a classical smoothness regularizer on the coil maps. The resulting procedure runs quickly, remains stable when contrast, anatomy, or sampling trajectory varies, and benefits from an upgraded parameterization of the prior that improves both standalone denoising and end-to-end MRI reconstruction.

What carries the argument

The product-of-Gaussian-mixture diffusion model used as the image prior, combined with a smoothness prior on the coil sensitivities, in a joint nonlinear optimization.

If this is right

Reconstruction no longer requires a separate offline step to estimate coil sensitivities.
The same trained prior remains effective under shifts in image contrast or anatomy.
Changing k-space trajectories can be accommodated without retraining the image model.
A more expressive parameterization of the diffusion prior directly raises quality in both denoising and full reconstruction.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

The joint formulation could be adapted to other multi-parameter inverse problems such as joint estimation of motion or field inhomogeneity maps.
Lower parameter count may allow the model to run on resource-limited hardware for on-site or portable scanner reconstruction.
If the prior generalizes further, it could support accelerated dynamic or multi-contrast acquisitions by reducing the need for per-sequence retraining.

Load-bearing premise

The chosen diffusion model supplies a sufficiently general prior for MRI images across varying contrasts and anatomies, while a basic smoothness penalty is enough to constrain the coil sensitivities without creating artifacts.

What would settle it

Systematic degradation or visible artifacts when the method is applied to a held-out dataset that introduces an unseen contrast, anatomical region, or non-Cartesian k-space trajectory not represented in training.

read the original abstract

Recently, diffusion models have attracted considerable attention for magnetic resonance image reconstruction due to their high sample quality. However, most existing methods rely on large networks with opaque time-conditioning mechanisms, and require offline coil sensitivity estimation. This results in limited interpretability of the reconstruction process and reduced flexibility in the acquisition setup. To address these limitations, we jointly reconstruct the image and the coil sensitivities by combining the parameter-efficient product-of-Gaussian-mixture diffusion model as an image prior with a classical smoothness prior on the coil sensitivities. The proposed method is fast and robust to both contrast and anatomical distribution shifts as well as changing k-space trajectories. Finally, we propose a more expressive parameterization of the image prior which improves results in denoising and magnetic resonance image reconstruction.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

This combines a product-of-Gaussian-mixture diffusion prior with joint coil sensitivity estimation to make MRI reconstruction more flexible and less reliant on large networks.

read the letter

The main point is that this method jointly reconstructs the image and coil sensitivities by pairing a product-of-Gaussian-mixture diffusion model as the image prior with a classical smoothness prior on the coils, plus a tweak to make the image prior more expressive. It positions itself as faster and more robust to changes in contrast, anatomy, and k-space trajectories than typical diffusion approaches that need big time-conditioned networks and separate coil estimation upfront. That joint setup and the parameter-efficient prior are the concrete advances here. The integration of the diffusion component with a simple classical term lets the method avoid some of the opacity and preprocessing steps that limit other diffusion MRI work, and the claim of handling distribution shifts without retraining is the practical hook if the results hold. The parameterization change is presented as helping both standalone denoising and the full reconstruction task, which shows they checked it on related problems. The soft spot is that the robustness to shifts rests on the mixture model generalizing as a prior and the smoothness assumption not creating artifacts; without detailed experiments on varied datasets and strong baselines, those claims stay hard to weigh. The math itself looks like a direct adaptation rather than a deep reformulation, so the value sits in the application and the joint formulation. This is for people working on generative priors for medical inverse problems or accelerated MRI. A reader who wants alternatives to heavy diffusion networks would get something concrete from the joint aspect and the efficiency angle. It has enough new framing and a clear target problem to deserve a serious referee who can check the experiments and the shift tests.

Referee Report

1 major / 0 minor

Summary. The manuscript proposes a joint nonlinear MRI reconstruction approach that estimates both the image and coil sensitivities simultaneously. It uses a parameter-efficient product-of-Gaussian-mixture diffusion model as the image prior combined with a classical smoothness prior on the coil sensitivities. The work claims that the resulting method is fast, robust to contrast and anatomical distribution shifts as well as changes in k-space trajectories, and that a more expressive parameterization of the image prior further improves results in both denoising and MRI reconstruction.

Significance. If the empirical validation supports the claims, the contribution would lie in providing a more interpretable and computationally lighter diffusion-based prior for MRI that enables joint coil-image estimation without offline sensitivity mapping. This could increase flexibility for varying acquisition protocols and improve generalization, addressing limitations of large time-conditioned networks. The product-of-Gaussian-mixture construction and the proposed expressive parameterization represent potentially useful modeling advances if they demonstrably outperform existing priors under distribution shift.

major comments (1)

Abstract: The abstract states that the method 'is fast and robust to both contrast and anatomical distribution shifts as well as changing k-space trajectories' and that the expressive parameterization 'improves results in denoising and magnetic resonance image reconstruction,' yet supplies no quantitative metrics, baseline comparisons, error bars, or experimental protocols. This absence prevents assessment of whether the central claims are supported by data.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for their review and the opportunity to respond. We address the major comment on the abstract below.

read point-by-point responses

Referee: Abstract: The abstract states that the method 'is fast and robust to both contrast and anatomical distribution shifts as well as changing k-space trajectories' and that the expressive parameterization 'improves results in denoising and magnetic resonance image reconstruction,' yet supplies no quantitative metrics, baseline comparisons, error bars, or experimental protocols. This absence prevents assessment of whether the central claims are supported by data.

Authors: We acknowledge that the abstract, as a concise summary, does not include quantitative metrics or experimental details. The manuscript provides these in the Experiments section, with baseline comparisons, error bars, and protocols demonstrating robustness to contrast/anatomical shifts and trajectory changes, plus improvements from the expressive parameterization in both denoising and reconstruction tasks. To address the concern, we will revise the abstract to incorporate key quantitative highlights supporting the claims. revision: yes

Circularity Check

0 steps flagged

No significant circularity

full rationale

The paper presents a joint reconstruction method that combines a parameter-efficient product-of-Gaussian-mixture diffusion model as an image prior with a classical smoothness prior on coil sensitivities, plus a proposed more expressive parameterization. No load-bearing derivation step is shown to reduce by construction to fitted inputs, self-definitions, or self-citation chains; the abstract and description frame the work as an empirical combination of priors with robustness claims that rest on external validation rather than tautological equivalence. The central claims concern practical performance across shifts and trajectories, which are not internally forced by the method's own definitions.

Axiom & Free-Parameter Ledger

1 free parameters · 2 axioms · 0 invented entities

The central claim rests on the effectiveness of the product-of-Gaussian-mixture diffusion model as an image prior and the adequacy of a smoothness prior for coil sensitivities; these are domain assumptions introduced to enable the joint reconstruction but lack independent verification in the abstract.

free parameters (1)

Gaussian mixture parameters
The product-of-Gaussian-mixture model is described as parameter-efficient, implying parameters that are learned or set during training or inference, though exact count and fitting procedure are not specified in the abstract.

axioms (2)

domain assumption Product-of-Gaussian-mixture diffusion model is an effective prior for MRI images
Invoked directly as the image prior to enable joint reconstruction and robustness to shifts.
domain assumption Classical smoothness prior sufficiently models coil sensitivities
Used to jointly estimate coil sensitivities without offline calibration.

pith-pipeline@v0.9.0 · 5419 in / 1440 out tokens · 59631 ms · 2026-05-12T05:12:17.888827+00:00 · methodology

discussion (0)

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

IndisputableMonolith/Cost/FunctionalEquation.lean (Jcost, washburn_uniqueness_aczel) reality_from_one_distinction unclear

?

unclear
Relation between the paper passage and the cited Recognition theorem.

p_θ(x,t)∝∏_l∏_k ψ_k((K_k x)_l,w_k,t) with ψ_k a 1-D Gaussian mixture whose variance evolves as σ_k²(t)=σ₀²+ν_k²·2t (or learned τ_θ(t))

What do these tags mean?

matches: The paper's claim is directly supported by a theorem in the formal canon.
supports: The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends: The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses: The paper appears to rely on the theorem as machinery.
contradicts: The paper's claim conflicts with a theorem or certificate in the canon.
unclear: Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

79 extracted references · 79 canonical work pages

[1]

Magnetic Resonance in Medicine42(5), 952–962 (1999)

Pruessmann, K.P., Weiger, M., Scheidegger, M.B., Boesiger, P.: Sense: Sensitivity encoding for fast mri. Magnetic Resonance in Medicine42(5), 952–962 (1999)

work page 1999
[2]

Magnetic Resonance in Medicine47(6), 1202–1210 (2002)

Griswold, M.A., Jakob, P.M., Heidemann, R.M., Nittka, M., Jellus, V., Wang, J., Kiefer, B., Haase, A.: Generalized autocalibrating partially parallel acquisitions (grappa). Magnetic Resonance in Medicine47(6), 1202–1210 (2002)

work page 2002
[3]

iterative image reconstruction using a total variation constraint

Block, K.T., Uecker, M., Frahm, J.: Undersampled radial mri with multiple coils. iterative image reconstruction using a total variation constraint. Magnetic Resonance in Medicine57(6), 1086–1098 (2007)

work page 2007
[4]

Magnetic Resonance in Medicine65(2), 480–491 (2011)

Knoll, F., Bredies, K., Pock, T., Stollberger, R.: Second order total generalized variation (tgv) for mri. Magnetic Resonance in Medicine65(2), 480–491 (2011)

work page 2011
[5]

Magnetic Resonance in Medicine79(6), 3055–3071 (2018)

Hammernik, K., Klatzer, T., Kobler, E., Recht, M.P., Sodickson, D.K., Pock, T., Knoll, F.: Learning a variational network for reconstruction of accelerated mri data. Magnetic Resonance in Medicine79(6), 3055–3071 (2018)

work page 2018
[6]

IEEE transactions on medical imagingPP(2023)

Zach, M., Knoll, F., Pock, T.: Stable deep mri reconstruction using generative priors. IEEE transactions on medical imagingPP(2023)

work page 2023
[7]

Journal of Magnetic Resonance Imaging36(1), 55–72 (2012)

Deshmane, A., Gulani, V., Griswold, M.A., Seiberlich, N.: Parallel mr imaging. Journal of Magnetic Resonance Imaging36(1), 55–72 (2012)

work page 2012
[8]

Progress in Nuclear Magnetic Resonance Spectroscopy101, 71–95 (2017)

Hamilton, J., Franson, D., Seiberlich, N.: Recent advances in parallel imaging for mri. Progress in Nuclear Magnetic Resonance Spectroscopy101, 71–95 (2017)

work page 2017
[9]

Applied Magnetic Resonance47(5), 487–498 (2016)

Irfan, A.S., Nisar, A., Shahzad, H., Omer, H.: Sensitivity maps estimation using eigenvalues in sense reconstruction. Applied Magnetic Resonance47(5), 487–498 (2016)

work page 2016
[10]

In: Proc

Ballester, M.G., Machida, Y., Kassai, Y., Hamamura, Y., Sugimoto, H.: Robust estimation of coil sensitivities for rf subencoding acquisition techniques. In: Proc. 9th Ann. Meeting ISMRM, p. 799 (2001)

work page 2001
[11]

Lin, F.-H., Wald, L., Belliveau, J.: Estimation of coil sensitivity map and correction of surface coil magnetic resonance images using wavelet decomposition (2001)

work page 2001
[12]

Human Brain Mapping19(2), 96–111 (2003)

Lin, F.-H., Chen, Y.-J., Belliveau, J.W., Wald, L.L.: A wavelet-based approx- imation of surface coil sensitivity profiles for correction of image intensity inhomogeneity and parallel imaging reconstruction. Human Brain Mapping19(2), 96–111 (2003)

work page 2003
[13]

Magnetic Resonance Imaging32(10), 1365–1376 (2014) 42

Ling, Q., Li, Z., Song, K., Li, F.: An iterative method for coil sensitivity estimation in multi-coil mri systems. Magnetic Resonance Imaging32(10), 1365–1376 (2014) 42

work page 2014
[14]

Magnetic Resonance in Medicine71(3), 990–1001 (2014)

Uecker, M., Lai, P., Murphy, M.J., Virtue, P., Elad, M., Pauly, J.M., Vasanawala, S.S., Lustig, M.: Espirit an eigenvalue approach to autocalibrating parallel mri: Where sense meets grappa. Magnetic Resonance in Medicine71(3), 990–1001 (2014)

work page 2014
[15]

Magnetic Resonance in Medicine57(6), 1196–1202 (2007)

Ying, L., Sheng, J.: Joint image reconstruction and sensitivity estimation in sense (jsense). Magnetic Resonance in Medicine57(6), 1196–1202 (2007)

work page 2007
[16]

Mathematical Methods in the Applied Sciences 30(12), 1437–1451 (2007)

Bauer, F., Kannengiesser, S.: An alternative approach to the image reconstruction for parallel data acquisition in mri. Mathematical Methods in the Applied Sciences 30(12), 1437–1451 (2007)

work page 2007
[18]

Magnetic Resonance in Medicine67(1), 34–41 (2012)

Knoll, F., Clason, C., Bredies, K., Uecker, M., Stollberger, R.: Parallel imaging with nonlinear reconstruction using variational penalties. Magnetic Resonance in Medicine67(1), 34–41 (2012)

work page 2012
[19]

Uecker, M., Block, K.T., Frahm, J.: Nonlinear inversion with l 1-wavelet regularization – application to autocalibrated parallel imaging. (2007)

work page 2007
[20]

SIAM Journal on Imaging Sciences3(3), 492–526 (2010)

Bredies, K., Kunisch, K., Pock, T.: Total generalized variation. SIAM Journal on Imaging Sciences3(3), 492–526 (2010)

work page 2010
[21]

Concepts in Magnetic Resonance Part A40A(6), 269–280 (2012)

Majumdar, A., Ward, R.K.: Iterative estimation of mri sensitivity maps and image based on sense reconstruction method (isense). Concepts in Magnetic Resonance Part A40A(6), 269–280 (2012)

work page 2012
[22]

IEEE Transactions on Medical Imaging39(2020)

Han, Y., Sunwoo, L., Ye, J.: K-space deep learning for accelerated mri. IEEE Transactions on Medical Imaging39(2020)

work page 2020
[23]

Zbontar, J., Knoll, F., Sriram, A., Muckley, M., Bruno, M., Defazio, A., Parente, M., Geras, K.J., Katsnelson, J., Chandarana, H., Zhang, Z., Drozdzal, M., Romero, A., Rabbat, M.G., Vincent, P., Pinkerton, J., Wang, D., Yakubova, N., Owens, E., Zitnick, C.L., Recht, M.P., Sodickson, D.K., Lui, Y.W.: fastmri: An open dataset and benchmarks for accelerated ...

work page 2018
[24]

IEEE Transactions on Medical Imaging38, 394–405 (2017)

Aggarwal, H.K., Mani, M.P., Jacob, M.: Modl: Model-based deep learning archi- tecture for inverse problems. IEEE Transactions on Medical Imaging38, 394–405 (2017)

work page 2017
[25]

Medical image computing and computer-assisted intervention : MICCAI International Conference on Medical Image Computing and Computer-Assisted Intervention12906, 350–360 (2021) 43

Arvinte, M., Vishwanath, S., Tewfik, A.H., Tamir, J.I.: Deep j-sense: Accelerated mri reconstruction via unrolled alternating optimization. Medical image computing and computer-assisted intervention : MICCAI International Conference on Medical Image Computing and Computer-Assisted Intervention12906, 350–360 (2021) 43

work page 2021
[26]

In: Wavelets and Sparsity XVIII (2019)

Narnhofer, D., Hammernik, K., Knoll, F., Pock, T.: Inverse gans for accelerated mri reconstruction. In: Wavelets and Sparsity XVIII (2019). SPIE Optics+Photonics 2019 ; Conference year: 23-08-2019 Through 27-08-2019

work page 2019
[27]

Advances in Neural Information Processing Systems34, 14938–14954 (2021)

Jalal, A., Arvinte, M., Daras, G., Price, E., Dimakis, A.G., Tamir, J.: Robust com- pressed sensing MRI with deep generative priors. Advances in Neural Information Processing Systems34, 14938–14954 (2021)

work page 2021
[28]

Medical Image Analysis80, 102479 (2022)

Chung, H., Ye, J.C.: Score-based diffusion models for accelerated mri. Medical Image Analysis80, 102479 (2022)

work page 2022
[29]

Magnetic Resonance in Medicine90(1), 295–311 (2023)

Luo, G., Blumenthal, M., Heide, M., Uecker, M.: Bayesian mri reconstruction with joint uncertainty estimation using diffusion models. Magnetic Resonance in Medicine90(1), 295–311 (2023)

work page 2023
[30]

NMR in Biomedicine 36(3), 4848 (2023)

Guan, Y., Tu, Z., Wang, S., Wang, Y., Liu, Q., Liang, D.: Magnetic resonance imaging reconstruction using a deep energy-based model. NMR in Biomedicine 36(3), 4848 (2023)

work page 2023
[31]

In: Navab, N., Hornegger, J., Wells, W.M., Frangi, A.F

Ronneberger, O., Fischer, P., Brox, T.: U-net: Convolutional networks for biomed- ical image segmentation. In: Navab, N., Hornegger, J., Wells, W.M., Frangi, A.F. (eds.) Medical Image Computing and Computer-Assisted Intervention – MICCAI 2015, pp. 234–241. Springer, Cham (2015)

work page 2015
[32]

Salimans, T., Ho, J.: Should EBMs model the energy or the score? In: Energy Based Models Workshop - ICLR 2021 (2021)

work page 2021
[33]

In: Proceedings of the 34th International Conference on Neural Information Processing Systems (2020)

Tancik, M., Srinivasan, P.P., Mildenhall, B., Fridovich-Keil, S., Raghavan, N., Singhal, U., Ramamoorthi, R., Barron, J.T., Ng, R.: Fourier features let networks learn high frequency functions in low dimensional domains. In: Proceedings of the 34th International Conference on Neural Information Processing Systems (2020)

work page 2020
[34]

In: Proceedings of the 31st International Conference on Neural Information Processing Systems (2017)

Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A.N., Kaiser, L., Polosukhin, I.: Attention is all you need. In: Proceedings of the 31st International Conference on Neural Information Processing Systems (2017)

work page 2017
[35]

Erlacher, M., Zach, M.: Joint Non-Linear MRI Inversion with Diffusion Priors (2023)

work page 2023
[36]

Journal of Mathematical Imaging and Vision66, 1–25 (2024)

Zach, M., Kobler, E., Chambolle, A., Pock, T.: Product of gaussian mixture diffusion models. Journal of Mathematical Imaging and Vision66, 1–25 (2024)

work page 2024
[37]

In: 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR’05), vol

Roth, S., Black, M.J.: Fields of experts: a framework for learning image priors. In: 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR’05), vol. 2 (2005)

work page 2005
[38]

In: NeurIPS 2023 Workshop on Deep Learning and Inverse Problems (2023)

Ozturkler, B., Mardani, M., Vahdat, A., Kautz, J., Pauly, J.M.: Regularization by 44 denoising diffusion process for MRI reconstruction. In: NeurIPS 2023 Workshop on Deep Learning and Inverse Problems (2023)

work page 2023
[39]

Hu, Y., Peng, A., Gan, W., Kamilov, U.S.: ADOBI: Adaptive Diffusion Bridge For Blind Inverse Problems with Application to MRI Reconstruction (2024)

work page 2024
[40]

Journal of Magnetic Resonance163(1), 121–123 (2003)

Larsson, E.G., Erdogmus, D., Yan, R., Principe, J.C., Fitzsimmons, J.R.: Snr- optimality of sum-of-squares reconstruction for phased-array magnetic resonance imaging. Journal of Magnetic Resonance163(1), 121–123 (2003)

work page 2003
[41]

IEEE/CVF Conference on Computer Vision and Pattern Recognition (2023)

Chung, H., Kim, J., Kim, S., Ye, J.C.: Parallel diffusion models of operator and image for blind inverse problems. IEEE/CVF Conference on Computer Vision and Pattern Recognition (2023)

work page 2023
[42]

In: Proceedings of Medical Image Computing and Computer Assisted Intervention – MICCAI 2024 (2024)

Li, X., Yang, Y., Zheng, H., Xu, Z.: Blind proximal diffusion model for joint image and sensitivity estimation in parallel mri. In: Proceedings of Medical Image Computing and Computer Assisted Intervention – MICCAI 2024 (2024)

work page 2024
[43]

Mathematical Programming146(1), 459–494 (2014)

Bolte, J., Sabach, S., Teboulle, M.: Proximal alternating linearized minimization for nonconvex and nonsmooth problems. Mathematical Programming146(1), 459–494 (2014)

work page 2014
[44]

In: 2010 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, pp

Schmidt, U., Gao, Q., Roth, S.: A generative perspective on mrfs in low-level vision. In: 2010 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, pp. 1751–1758 (2010)

work page 2010
[45]

In: Proceedings of the 40th International Conference on Machine Learning, pp

Murata, N., Saito, K., Lai, C.-H., Takida, Y., Uesaka, T., Mitsufuji, Y., Ermon, S.: GibbsDDRM: A partially collapsed Gibbs sampler for solving blind inverse prob- lems with denoising diffusion restoration. In: Proceedings of the 40th International Conference on Machine Learning, pp. 25501–25522 (2023)

work page 2023
[46]

Chung, H., Kim, J., Ye, J.C.: Diffusion models for inverse problems (2025)

work page 2025
[47]

Daras, G., Chung, H., Lai, C.-H., Mitsufuji, Y., Ye, J.C., Milanfar, P., Dimakis, A.G., Delbracio, M.: A Survey on Diffusion Models for Inverse Problems (2024)

work page 2024
[48]

In: Bubba, T.A., Gaburro, R., Gazzola, S., Papafitsoros, K., Pereyra, M., Sch¨ onlieb, C.-B

Nagler, L., Zach, M., Pock, T.: Product of gaussian mixture diffusion model for non- linear mri inversion. In: Bubba, T.A., Gaburro, R., Gazzola, S., Papafitsoros, K., Pereyra, M., Sch¨ onlieb, C.-B. (eds.) Scale Space and Variational Methods in Computer Vision, pp. 146–158. Springer, Cham (2025)

work page 2025
[49]

Magnetic Resonance in Medicine34(6), 910–914 (1995)

Gudbjartsson, H., Patz, S.: The rician distribution of noisy mri data. Magnetic Resonance in Medicine34(6), 910–914 (1995)

work page 1995
[50]

SIAM Journal on Imaging Sciences12(2019)

Pereyra, M.: Revisiting maximum-a-posteriori estimation in log-concave models. SIAM Journal on Imaging Sciences12(2019)

work page 2019
[51]

In: Ghanem, 45 R., Higdon, D., Owhadi, H

Dashti, M., Stuart, A.M.: The bayesian approach to inverse problems. In: Ghanem, 45 R., Higdon, D., Owhadi, H. (eds.) Handbook of Uncertainty Quantification, pp. 311–428. Springer, Cham (2017)

work page 2017
[52]

In: Pattern Recognition: 46th DAGM German Conference, DAGM GCPR 2024, Munich, Germany, September 10–13, 2024, Proceedings, Part I, pp

Glaszner, L., Zach, M.: Bigger isn’t always better: Towards a general prior for medical image reconstruction. In: Pattern Recognition: 46th DAGM German Conference, DAGM GCPR 2024, Munich, Germany, September 10–13, 2024, Proceedings, Part I, pp. 275–291 (2025)

work page 2024
[53]

In: Advances in Neural Information Processing Systems, pp

Song, Y., Ermon, S.: Generative modeling by estimating gradients of the data distribution. In: Advances in Neural Information Processing Systems, pp. 11895– 11907 (2019)

work page 2019
[54]

Habring, A., Holler, M., Pock, T., Zach, M.: Energy-based models for inverse imaging problems (2025)

work page 2025
[55]

In: International Conference on Learning Representations (2021)

Song, Y., Sohl-Dickstein, J., Kingma, D.P., Kumar, A., Ermon, S., Poole, B.: Score-based generative modeling through stochastic differential equations. In: International Conference on Learning Representations (2021)

work page 2021
[56]

Cambridge University Press, Cambridge (2019)

S¨ arkk¨ a, S., Solin, A.: Applied Stochastic Differential Equations. Cambridge University Press, Cambridge (2019)

work page 2019
[57]

In: The Fokker-Planck Equation: Methods of Solution and Applications, pp

Risken, H.: Fokker-planck equation. In: The Fokker-Planck Equation: Methods of Solution and Applications, pp. 63–95. Springer, Berlin, Heidelberg (1996)

work page 1996
[58]

Yanovsky, I.: Partial differential equations: Graduate level problems and solutions (2005)

work page 2005
[59]

convergence to the gaussian (2017)

V’azquez, J.L.: Asymptotic behaviour methods for the heat equation. convergence to the gaussian (2017)

work page 2017
[60]

Stochastic Processes and their Applications12(3), 313–326 (1982)

Anderson, B.D.O.: Reverse-time diffusion equation models. Stochastic Processes and their Applications12(3), 313–326 (1982)

work page 1982
[61]

Journal of the American Statistical Association106(496), 1602–1614 (2011)

Efron, B.: Tweedie’s formula and selection bias. Journal of the American Statistical Association106(496), 1602–1614 (2011)

work page 2011
[62]

Robbins, H.E.: An empirical bayes approach to statistics. (1956)

work page 1956
[63]

Neural Computation23, 1661–1674 (2011)

Vincent, P.: A connection between score matching and denoising autoencoders. Neural Computation23, 1661–1674 (2011)

work page 2011
[64]

In: Neural Information Processing Systems (2021)

Meng, C., Song, Y., Li, W., Ermon, S.: Estimating high order gradients of the data distribution by denoising. In: Neural Information Processing Systems (2021)

work page 2021
[65]

In: The Eleventh International Conference on Learning Representations (2023) 46

Chung, H., Kim, J., Mccann, M.T., Klasky, M.L., Ye, J.C.: Diffusion posterior sampling for general noisy inverse problems. In: The Eleventh International Conference on Learning Representations (2023) 46

work page 2023
[66]

Prentice-Hall, Inc., USA (1996)

Proakis, J.G., Manolakis, D.G.: Digital Signal Processing (3rd Ed.): Principles, Algorithms, and Applications. Prentice-Hall, Inc., USA (1996)

work page 1996
[67]

IEEE Transactions on Image Processing22(5), 2056–2065 (2013)

Lim, W.-Q.: Nonseparable shearlet transform. IEEE Transactions on Image Processing22(5), 2056–2065 (2013)

work page 2056
[68]

Magnetic Resonance in Medicine60(3), 674–682 (2008)

Uecker, M., Hohage, T., Block, K.T., Frahm, J.: Image reconstruction by regular- ized nonlinear inversion—joint estimation of coil sensitivities and image content. Magnetic Resonance in Medicine60(3), 674–682 (2008)

work page 2008
[69]

Cambridge University Press, USA (1992)

Press, W.H., Teukolsky, S.A., Vetterling, W.T., Flannery, B.P.: Numerical Recipes in C (2nd Ed.): the Art of Scientific Computing. Cambridge University Press, USA (1992)

work page 1992
[70]

Radiology: Artificial Intelligence2, 190007 (2020)

Knoll, F., Zbontar, J., Sriram, A., Muckley, M., Bruno, M., Defazio, A., Parente, M., Geras, K., Katsnelson, J., Chandarana, H., Zhang, Z., Drozdzalv, M., Romero, A., Rabbat, M., Vincent, P., Pinkerton, J., Wang, D., Yakubova, N., Owens, E., Lui, Y.: fastmri: A publicly available raw k-space and dicom dataset of knee images for accelerated mr image recons...

work page 2020
[71]

2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 12403–12412 (2021)

Chung, H., Sim, B., Ye, J.-C.: Come-closer-diffuse-faster: Accelerating condi- tional diffusion models for inverse problems through stochastic contraction. 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 12403–12412 (2021)

work page 2022
[72]

IEEE Transactions on Image Processing23(3), 1060–1072 (2014)

Chen, Y., Ranftl, R., Pock, T.: Insights into analysis operator learning: From patch-based sparse models to higher order mrfs. IEEE Transactions on Image Processing23(3), 1060–1072 (2014)

work page 2014
[73]

Magnetic Resonance in Medicine42(1999)

Ny´ ul, L.G., Udupa, J.K.: On standardizing the mr image intensity scale. Magnetic Resonance in Medicine42(1999)

work page 1999
[74]

Conference on Neural Information Processing Systems (2020)

Zhuang, J., Tang, T., Ding, Y., Tatikonda, S., Dvornek, N., Papademetris, X., Duncan, J.: Adabelief optimizer: Adapting stepsizes by the belief in observed gradients. Conference on Neural Information Processing Systems (2020)

work page 2020
[75]

Pattern Analysis and Machine Intelligence, IEEE Transactions on19, 1236–1250 (1997)

Zhu, S., Mumford, D.: Prior learning and gibbs reaction-diffusion. Pattern Analysis and Machine Intelligence, IEEE Transactions on19, 1236–1250 (1997)

work page 1997
[76]

IEEE Transactions on Pattern Analysis and Machine Intelligence39(6), 1256–1272 (2017)

Chen, Y., Pock, T.: Trainable nonlinear reaction diffusion: A flexible framework for fast and effective image restoration. IEEE Transactions on Pattern Analysis and Machine Intelligence39(6), 1256–1272 (2017)

work page 2017
[77]

In: Medical Image Computing and Computer Assisted Intervention – MICCAI 47 2020 (2020)

Sriram, A., Zbontar, J., Murrell, T., Defazio, A., Zitnick, C.L., Yakubova, N., Knoll, F., Johnson, P.: End-to-end variational networks for accelerated mri reconstruction. In: Medical Image Computing and Computer Assisted Intervention – MICCAI 47 2020 (2020)

work page 2020
[78]

In: Proceedings of the 34th International Conference on Neural Information Processing Systems (2020)

Song, Y., Ermon, S.: Improved techniques for training score-based generative models. In: Proceedings of the 34th International Conference on Neural Information Processing Systems (2020)

work page 2020
[79]

In: IEEE Conference on Computer Vision and Pattern Recognition, pp

Zhang, K., Zuo, W., Gu, S., Zhang, L.: Learning deep cnn denoiser prior for image restoration. In: IEEE Conference on Computer Vision and Pattern Recognition, pp. 3929–3938 (2017)

work page 2017
[80]

In: Bengio, Y., LeCun, Y

Kingma, D.P., Ba, J.: Adam: A method for stochastic optimization. In: Bengio, Y., LeCun, Y. (eds.) ICLR (Poster) (2015) 48

work page 2015

[1] [1]

Magnetic Resonance in Medicine42(5), 952–962 (1999)

Pruessmann, K.P., Weiger, M., Scheidegger, M.B., Boesiger, P.: Sense: Sensitivity encoding for fast mri. Magnetic Resonance in Medicine42(5), 952–962 (1999)

work page 1999

[2] [2]

Magnetic Resonance in Medicine47(6), 1202–1210 (2002)

Griswold, M.A., Jakob, P.M., Heidemann, R.M., Nittka, M., Jellus, V., Wang, J., Kiefer, B., Haase, A.: Generalized autocalibrating partially parallel acquisitions (grappa). Magnetic Resonance in Medicine47(6), 1202–1210 (2002)

work page 2002

[3] [3]

iterative image reconstruction using a total variation constraint

Block, K.T., Uecker, M., Frahm, J.: Undersampled radial mri with multiple coils. iterative image reconstruction using a total variation constraint. Magnetic Resonance in Medicine57(6), 1086–1098 (2007)

work page 2007

[4] [4]

Magnetic Resonance in Medicine65(2), 480–491 (2011)

Knoll, F., Bredies, K., Pock, T., Stollberger, R.: Second order total generalized variation (tgv) for mri. Magnetic Resonance in Medicine65(2), 480–491 (2011)

work page 2011

[5] [5]

Magnetic Resonance in Medicine79(6), 3055–3071 (2018)

Hammernik, K., Klatzer, T., Kobler, E., Recht, M.P., Sodickson, D.K., Pock, T., Knoll, F.: Learning a variational network for reconstruction of accelerated mri data. Magnetic Resonance in Medicine79(6), 3055–3071 (2018)

work page 2018

[6] [6]

IEEE transactions on medical imagingPP(2023)

Zach, M., Knoll, F., Pock, T.: Stable deep mri reconstruction using generative priors. IEEE transactions on medical imagingPP(2023)

work page 2023

[7] [7]

Journal of Magnetic Resonance Imaging36(1), 55–72 (2012)

Deshmane, A., Gulani, V., Griswold, M.A., Seiberlich, N.: Parallel mr imaging. Journal of Magnetic Resonance Imaging36(1), 55–72 (2012)

work page 2012

[8] [8]

Progress in Nuclear Magnetic Resonance Spectroscopy101, 71–95 (2017)

Hamilton, J., Franson, D., Seiberlich, N.: Recent advances in parallel imaging for mri. Progress in Nuclear Magnetic Resonance Spectroscopy101, 71–95 (2017)

work page 2017

[9] [9]

Applied Magnetic Resonance47(5), 487–498 (2016)

Irfan, A.S., Nisar, A., Shahzad, H., Omer, H.: Sensitivity maps estimation using eigenvalues in sense reconstruction. Applied Magnetic Resonance47(5), 487–498 (2016)

work page 2016

[10] [10]

In: Proc

Ballester, M.G., Machida, Y., Kassai, Y., Hamamura, Y., Sugimoto, H.: Robust estimation of coil sensitivities for rf subencoding acquisition techniques. In: Proc. 9th Ann. Meeting ISMRM, p. 799 (2001)

work page 2001

[11] [11]

Lin, F.-H., Wald, L., Belliveau, J.: Estimation of coil sensitivity map and correction of surface coil magnetic resonance images using wavelet decomposition (2001)

work page 2001

[12] [12]

Human Brain Mapping19(2), 96–111 (2003)

Lin, F.-H., Chen, Y.-J., Belliveau, J.W., Wald, L.L.: A wavelet-based approx- imation of surface coil sensitivity profiles for correction of image intensity inhomogeneity and parallel imaging reconstruction. Human Brain Mapping19(2), 96–111 (2003)

work page 2003

[13] [13]

Magnetic Resonance Imaging32(10), 1365–1376 (2014) 42

Ling, Q., Li, Z., Song, K., Li, F.: An iterative method for coil sensitivity estimation in multi-coil mri systems. Magnetic Resonance Imaging32(10), 1365–1376 (2014) 42

work page 2014

[14] [14]

Magnetic Resonance in Medicine71(3), 990–1001 (2014)

Uecker, M., Lai, P., Murphy, M.J., Virtue, P., Elad, M., Pauly, J.M., Vasanawala, S.S., Lustig, M.: Espirit an eigenvalue approach to autocalibrating parallel mri: Where sense meets grappa. Magnetic Resonance in Medicine71(3), 990–1001 (2014)

work page 2014

[15] [15]

Magnetic Resonance in Medicine57(6), 1196–1202 (2007)

Ying, L., Sheng, J.: Joint image reconstruction and sensitivity estimation in sense (jsense). Magnetic Resonance in Medicine57(6), 1196–1202 (2007)

work page 2007

[16] [16]

Mathematical Methods in the Applied Sciences 30(12), 1437–1451 (2007)

Bauer, F., Kannengiesser, S.: An alternative approach to the image reconstruction for parallel data acquisition in mri. Mathematical Methods in the Applied Sciences 30(12), 1437–1451 (2007)

work page 2007

[17] [18]

Magnetic Resonance in Medicine67(1), 34–41 (2012)

Knoll, F., Clason, C., Bredies, K., Uecker, M., Stollberger, R.: Parallel imaging with nonlinear reconstruction using variational penalties. Magnetic Resonance in Medicine67(1), 34–41 (2012)

work page 2012

[18] [19]

Uecker, M., Block, K.T., Frahm, J.: Nonlinear inversion with l 1-wavelet regularization – application to autocalibrated parallel imaging. (2007)

work page 2007

[19] [20]

SIAM Journal on Imaging Sciences3(3), 492–526 (2010)

Bredies, K., Kunisch, K., Pock, T.: Total generalized variation. SIAM Journal on Imaging Sciences3(3), 492–526 (2010)

work page 2010

[20] [21]

Concepts in Magnetic Resonance Part A40A(6), 269–280 (2012)

Majumdar, A., Ward, R.K.: Iterative estimation of mri sensitivity maps and image based on sense reconstruction method (isense). Concepts in Magnetic Resonance Part A40A(6), 269–280 (2012)

work page 2012

[21] [22]

IEEE Transactions on Medical Imaging39(2020)

Han, Y., Sunwoo, L., Ye, J.: K-space deep learning for accelerated mri. IEEE Transactions on Medical Imaging39(2020)

work page 2020

[22] [23]

Zbontar, J., Knoll, F., Sriram, A., Muckley, M., Bruno, M., Defazio, A., Parente, M., Geras, K.J., Katsnelson, J., Chandarana, H., Zhang, Z., Drozdzal, M., Romero, A., Rabbat, M.G., Vincent, P., Pinkerton, J., Wang, D., Yakubova, N., Owens, E., Zitnick, C.L., Recht, M.P., Sodickson, D.K., Lui, Y.W.: fastmri: An open dataset and benchmarks for accelerated ...

work page 2018

[23] [24]

IEEE Transactions on Medical Imaging38, 394–405 (2017)

Aggarwal, H.K., Mani, M.P., Jacob, M.: Modl: Model-based deep learning archi- tecture for inverse problems. IEEE Transactions on Medical Imaging38, 394–405 (2017)

work page 2017

[24] [25]

Medical image computing and computer-assisted intervention : MICCAI International Conference on Medical Image Computing and Computer-Assisted Intervention12906, 350–360 (2021) 43

Arvinte, M., Vishwanath, S., Tewfik, A.H., Tamir, J.I.: Deep j-sense: Accelerated mri reconstruction via unrolled alternating optimization. Medical image computing and computer-assisted intervention : MICCAI International Conference on Medical Image Computing and Computer-Assisted Intervention12906, 350–360 (2021) 43

work page 2021

[25] [26]

In: Wavelets and Sparsity XVIII (2019)

Narnhofer, D., Hammernik, K., Knoll, F., Pock, T.: Inverse gans for accelerated mri reconstruction. In: Wavelets and Sparsity XVIII (2019). SPIE Optics+Photonics 2019 ; Conference year: 23-08-2019 Through 27-08-2019

work page 2019

[26] [27]

Advances in Neural Information Processing Systems34, 14938–14954 (2021)

Jalal, A., Arvinte, M., Daras, G., Price, E., Dimakis, A.G., Tamir, J.: Robust com- pressed sensing MRI with deep generative priors. Advances in Neural Information Processing Systems34, 14938–14954 (2021)

work page 2021

[27] [28]

Medical Image Analysis80, 102479 (2022)

Chung, H., Ye, J.C.: Score-based diffusion models for accelerated mri. Medical Image Analysis80, 102479 (2022)

work page 2022

[28] [29]

Magnetic Resonance in Medicine90(1), 295–311 (2023)

Luo, G., Blumenthal, M., Heide, M., Uecker, M.: Bayesian mri reconstruction with joint uncertainty estimation using diffusion models. Magnetic Resonance in Medicine90(1), 295–311 (2023)

work page 2023

[29] [30]

NMR in Biomedicine 36(3), 4848 (2023)

Guan, Y., Tu, Z., Wang, S., Wang, Y., Liu, Q., Liang, D.: Magnetic resonance imaging reconstruction using a deep energy-based model. NMR in Biomedicine 36(3), 4848 (2023)

work page 2023

[30] [31]

In: Navab, N., Hornegger, J., Wells, W.M., Frangi, A.F

Ronneberger, O., Fischer, P., Brox, T.: U-net: Convolutional networks for biomed- ical image segmentation. In: Navab, N., Hornegger, J., Wells, W.M., Frangi, A.F. (eds.) Medical Image Computing and Computer-Assisted Intervention – MICCAI 2015, pp. 234–241. Springer, Cham (2015)

work page 2015

[31] [32]

Salimans, T., Ho, J.: Should EBMs model the energy or the score? In: Energy Based Models Workshop - ICLR 2021 (2021)

work page 2021

[32] [33]

In: Proceedings of the 34th International Conference on Neural Information Processing Systems (2020)

Tancik, M., Srinivasan, P.P., Mildenhall, B., Fridovich-Keil, S., Raghavan, N., Singhal, U., Ramamoorthi, R., Barron, J.T., Ng, R.: Fourier features let networks learn high frequency functions in low dimensional domains. In: Proceedings of the 34th International Conference on Neural Information Processing Systems (2020)

work page 2020

[33] [34]

In: Proceedings of the 31st International Conference on Neural Information Processing Systems (2017)

Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A.N., Kaiser, L., Polosukhin, I.: Attention is all you need. In: Proceedings of the 31st International Conference on Neural Information Processing Systems (2017)

work page 2017

[34] [35]

Erlacher, M., Zach, M.: Joint Non-Linear MRI Inversion with Diffusion Priors (2023)

work page 2023

[35] [36]

Journal of Mathematical Imaging and Vision66, 1–25 (2024)

Zach, M., Kobler, E., Chambolle, A., Pock, T.: Product of gaussian mixture diffusion models. Journal of Mathematical Imaging and Vision66, 1–25 (2024)

work page 2024

[36] [37]

In: 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR’05), vol

Roth, S., Black, M.J.: Fields of experts: a framework for learning image priors. In: 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR’05), vol. 2 (2005)

work page 2005

[37] [38]

In: NeurIPS 2023 Workshop on Deep Learning and Inverse Problems (2023)

Ozturkler, B., Mardani, M., Vahdat, A., Kautz, J., Pauly, J.M.: Regularization by 44 denoising diffusion process for MRI reconstruction. In: NeurIPS 2023 Workshop on Deep Learning and Inverse Problems (2023)

work page 2023

[38] [39]

Hu, Y., Peng, A., Gan, W., Kamilov, U.S.: ADOBI: Adaptive Diffusion Bridge For Blind Inverse Problems with Application to MRI Reconstruction (2024)

work page 2024

[39] [40]

Journal of Magnetic Resonance163(1), 121–123 (2003)

Larsson, E.G., Erdogmus, D., Yan, R., Principe, J.C., Fitzsimmons, J.R.: Snr- optimality of sum-of-squares reconstruction for phased-array magnetic resonance imaging. Journal of Magnetic Resonance163(1), 121–123 (2003)

work page 2003

[40] [41]

IEEE/CVF Conference on Computer Vision and Pattern Recognition (2023)

Chung, H., Kim, J., Kim, S., Ye, J.C.: Parallel diffusion models of operator and image for blind inverse problems. IEEE/CVF Conference on Computer Vision and Pattern Recognition (2023)

work page 2023

[41] [42]

In: Proceedings of Medical Image Computing and Computer Assisted Intervention – MICCAI 2024 (2024)

Li, X., Yang, Y., Zheng, H., Xu, Z.: Blind proximal diffusion model for joint image and sensitivity estimation in parallel mri. In: Proceedings of Medical Image Computing and Computer Assisted Intervention – MICCAI 2024 (2024)

work page 2024

[42] [43]

Mathematical Programming146(1), 459–494 (2014)

Bolte, J., Sabach, S., Teboulle, M.: Proximal alternating linearized minimization for nonconvex and nonsmooth problems. Mathematical Programming146(1), 459–494 (2014)

work page 2014

[43] [44]

In: 2010 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, pp

Schmidt, U., Gao, Q., Roth, S.: A generative perspective on mrfs in low-level vision. In: 2010 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, pp. 1751–1758 (2010)

work page 2010

[44] [45]

In: Proceedings of the 40th International Conference on Machine Learning, pp

Murata, N., Saito, K., Lai, C.-H., Takida, Y., Uesaka, T., Mitsufuji, Y., Ermon, S.: GibbsDDRM: A partially collapsed Gibbs sampler for solving blind inverse prob- lems with denoising diffusion restoration. In: Proceedings of the 40th International Conference on Machine Learning, pp. 25501–25522 (2023)

work page 2023

[45] [46]

Chung, H., Kim, J., Ye, J.C.: Diffusion models for inverse problems (2025)

work page 2025

[46] [47]

Daras, G., Chung, H., Lai, C.-H., Mitsufuji, Y., Ye, J.C., Milanfar, P., Dimakis, A.G., Delbracio, M.: A Survey on Diffusion Models for Inverse Problems (2024)

work page 2024

[47] [48]

In: Bubba, T.A., Gaburro, R., Gazzola, S., Papafitsoros, K., Pereyra, M., Sch¨ onlieb, C.-B

Nagler, L., Zach, M., Pock, T.: Product of gaussian mixture diffusion model for non- linear mri inversion. In: Bubba, T.A., Gaburro, R., Gazzola, S., Papafitsoros, K., Pereyra, M., Sch¨ onlieb, C.-B. (eds.) Scale Space and Variational Methods in Computer Vision, pp. 146–158. Springer, Cham (2025)

work page 2025

[48] [49]

Magnetic Resonance in Medicine34(6), 910–914 (1995)

Gudbjartsson, H., Patz, S.: The rician distribution of noisy mri data. Magnetic Resonance in Medicine34(6), 910–914 (1995)

work page 1995

[49] [50]

SIAM Journal on Imaging Sciences12(2019)

Pereyra, M.: Revisiting maximum-a-posteriori estimation in log-concave models. SIAM Journal on Imaging Sciences12(2019)

work page 2019

[50] [51]

In: Ghanem, 45 R., Higdon, D., Owhadi, H

Dashti, M., Stuart, A.M.: The bayesian approach to inverse problems. In: Ghanem, 45 R., Higdon, D., Owhadi, H. (eds.) Handbook of Uncertainty Quantification, pp. 311–428. Springer, Cham (2017)

work page 2017

[51] [52]

In: Pattern Recognition: 46th DAGM German Conference, DAGM GCPR 2024, Munich, Germany, September 10–13, 2024, Proceedings, Part I, pp

Glaszner, L., Zach, M.: Bigger isn’t always better: Towards a general prior for medical image reconstruction. In: Pattern Recognition: 46th DAGM German Conference, DAGM GCPR 2024, Munich, Germany, September 10–13, 2024, Proceedings, Part I, pp. 275–291 (2025)

work page 2024

[52] [53]

In: Advances in Neural Information Processing Systems, pp

Song, Y., Ermon, S.: Generative modeling by estimating gradients of the data distribution. In: Advances in Neural Information Processing Systems, pp. 11895– 11907 (2019)

work page 2019

[53] [54]

Habring, A., Holler, M., Pock, T., Zach, M.: Energy-based models for inverse imaging problems (2025)

work page 2025

[54] [55]

In: International Conference on Learning Representations (2021)

Song, Y., Sohl-Dickstein, J., Kingma, D.P., Kumar, A., Ermon, S., Poole, B.: Score-based generative modeling through stochastic differential equations. In: International Conference on Learning Representations (2021)

work page 2021

[55] [56]

Cambridge University Press, Cambridge (2019)

S¨ arkk¨ a, S., Solin, A.: Applied Stochastic Differential Equations. Cambridge University Press, Cambridge (2019)

work page 2019

[56] [57]

In: The Fokker-Planck Equation: Methods of Solution and Applications, pp

Risken, H.: Fokker-planck equation. In: The Fokker-Planck Equation: Methods of Solution and Applications, pp. 63–95. Springer, Berlin, Heidelberg (1996)

work page 1996

[57] [58]

Yanovsky, I.: Partial differential equations: Graduate level problems and solutions (2005)

work page 2005

[58] [59]

convergence to the gaussian (2017)

V’azquez, J.L.: Asymptotic behaviour methods for the heat equation. convergence to the gaussian (2017)

work page 2017

[59] [60]

Stochastic Processes and their Applications12(3), 313–326 (1982)

Anderson, B.D.O.: Reverse-time diffusion equation models. Stochastic Processes and their Applications12(3), 313–326 (1982)

work page 1982

[60] [61]

Journal of the American Statistical Association106(496), 1602–1614 (2011)

Efron, B.: Tweedie’s formula and selection bias. Journal of the American Statistical Association106(496), 1602–1614 (2011)

work page 2011

[61] [62]

Robbins, H.E.: An empirical bayes approach to statistics. (1956)

work page 1956

[62] [63]

Neural Computation23, 1661–1674 (2011)

Vincent, P.: A connection between score matching and denoising autoencoders. Neural Computation23, 1661–1674 (2011)

work page 2011

[63] [64]

In: Neural Information Processing Systems (2021)

Meng, C., Song, Y., Li, W., Ermon, S.: Estimating high order gradients of the data distribution by denoising. In: Neural Information Processing Systems (2021)

work page 2021

[64] [65]

In: The Eleventh International Conference on Learning Representations (2023) 46

Chung, H., Kim, J., Mccann, M.T., Klasky, M.L., Ye, J.C.: Diffusion posterior sampling for general noisy inverse problems. In: The Eleventh International Conference on Learning Representations (2023) 46

work page 2023

[65] [66]

Prentice-Hall, Inc., USA (1996)

Proakis, J.G., Manolakis, D.G.: Digital Signal Processing (3rd Ed.): Principles, Algorithms, and Applications. Prentice-Hall, Inc., USA (1996)

work page 1996

[66] [67]

IEEE Transactions on Image Processing22(5), 2056–2065 (2013)

Lim, W.-Q.: Nonseparable shearlet transform. IEEE Transactions on Image Processing22(5), 2056–2065 (2013)

work page 2056

[67] [68]

Magnetic Resonance in Medicine60(3), 674–682 (2008)

Uecker, M., Hohage, T., Block, K.T., Frahm, J.: Image reconstruction by regular- ized nonlinear inversion—joint estimation of coil sensitivities and image content. Magnetic Resonance in Medicine60(3), 674–682 (2008)

work page 2008

[68] [69]

Cambridge University Press, USA (1992)

Press, W.H., Teukolsky, S.A., Vetterling, W.T., Flannery, B.P.: Numerical Recipes in C (2nd Ed.): the Art of Scientific Computing. Cambridge University Press, USA (1992)

work page 1992

[69] [70]

Radiology: Artificial Intelligence2, 190007 (2020)

Knoll, F., Zbontar, J., Sriram, A., Muckley, M., Bruno, M., Defazio, A., Parente, M., Geras, K., Katsnelson, J., Chandarana, H., Zhang, Z., Drozdzalv, M., Romero, A., Rabbat, M., Vincent, P., Pinkerton, J., Wang, D., Yakubova, N., Owens, E., Lui, Y.: fastmri: A publicly available raw k-space and dicom dataset of knee images for accelerated mr image recons...

work page 2020

[70] [71]

2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 12403–12412 (2021)

Chung, H., Sim, B., Ye, J.-C.: Come-closer-diffuse-faster: Accelerating condi- tional diffusion models for inverse problems through stochastic contraction. 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 12403–12412 (2021)

work page 2022

[71] [72]

IEEE Transactions on Image Processing23(3), 1060–1072 (2014)

Chen, Y., Ranftl, R., Pock, T.: Insights into analysis operator learning: From patch-based sparse models to higher order mrfs. IEEE Transactions on Image Processing23(3), 1060–1072 (2014)

work page 2014

[72] [73]

Magnetic Resonance in Medicine42(1999)

Ny´ ul, L.G., Udupa, J.K.: On standardizing the mr image intensity scale. Magnetic Resonance in Medicine42(1999)

work page 1999

[73] [74]

Conference on Neural Information Processing Systems (2020)

Zhuang, J., Tang, T., Ding, Y., Tatikonda, S., Dvornek, N., Papademetris, X., Duncan, J.: Adabelief optimizer: Adapting stepsizes by the belief in observed gradients. Conference on Neural Information Processing Systems (2020)

work page 2020

[74] [75]

Pattern Analysis and Machine Intelligence, IEEE Transactions on19, 1236–1250 (1997)

Zhu, S., Mumford, D.: Prior learning and gibbs reaction-diffusion. Pattern Analysis and Machine Intelligence, IEEE Transactions on19, 1236–1250 (1997)

work page 1997

[75] [76]

IEEE Transactions on Pattern Analysis and Machine Intelligence39(6), 1256–1272 (2017)

Chen, Y., Pock, T.: Trainable nonlinear reaction diffusion: A flexible framework for fast and effective image restoration. IEEE Transactions on Pattern Analysis and Machine Intelligence39(6), 1256–1272 (2017)

work page 2017

[76] [77]

In: Medical Image Computing and Computer Assisted Intervention – MICCAI 47 2020 (2020)

Sriram, A., Zbontar, J., Murrell, T., Defazio, A., Zitnick, C.L., Yakubova, N., Knoll, F., Johnson, P.: End-to-end variational networks for accelerated mri reconstruction. In: Medical Image Computing and Computer Assisted Intervention – MICCAI 47 2020 (2020)

work page 2020

[77] [78]

In: Proceedings of the 34th International Conference on Neural Information Processing Systems (2020)

Song, Y., Ermon, S.: Improved techniques for training score-based generative models. In: Proceedings of the 34th International Conference on Neural Information Processing Systems (2020)

work page 2020

[78] [79]

In: IEEE Conference on Computer Vision and Pattern Recognition, pp

Zhang, K., Zuo, W., Gu, S., Zhang, L.: Learning deep cnn denoiser prior for image restoration. In: IEEE Conference on Computer Vision and Pattern Recognition, pp. 3929–3938 (2017)

work page 2017

[79] [80]

In: Bengio, Y., LeCun, Y

Kingma, D.P., Ba, J.: Adam: A method for stochastic optimization. In: Bengio, Y., LeCun, Y. (eds.) ICLR (Poster) (2015) 48

work page 2015