Pith. sign in

REVIEW 3 major objections 5 minor 74 references

Deep End-to-end Adaptive k-Space Sampling, Reconstruction, and Registration for Dynamic MRI

T0 review · 3 major / 5 minor · reviewed 2026-08-12 · deepseek-v4-flash

Pith's one-line read This paper claims that an end-to-end deep learning pipeline that jointly learns adaptive k-space sampling, image reconstruction, and deformable registration for dynamic MRI improves motion estimation over training each stage separately.

desk verdict Useful end-to-end dynamic MRI pipeline, but the registration claim rests on a metric that is also the training loss; needs an independent deformation-field evaluation before the gains can be believed. read the letter →

arxiv 2411.18249 v2 pith:5GFBRZIW submitted 2024-11-27 eess.IV cs.CVphysics.med-ph

classification eess.IVcs.CVphysics.med-ph
keywords dynamicMRIadaptivek-spacesamplingundersampledreconstructiondeformableimageregistrationmotionestimationend-to-enddeeplearningcardiaccineMR-guidedradiotherapy
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The paper is trying to establish that motion estimation from undersampled dynamic MRI is best done by one differentiable system that learns where to sample k-space, how to reconstruct images from those samples, and how to warp the reconstructed frames to a reference image, all trained together. It introduces a pipeline whose adaptive sampling module chooses case-specific k-space lines, whose reconstruction module is an unrolled optimization network, and whose registration module predicts deformation fields aligning the moving frames to a static reference. Controlled experiments on cardiac cine data and an unseen aorta dataset show that joint end-to-end training improves registration quality over training the same modules separately, and that learned phase-specific sampling helps motion estimation more than fixed sampling schemes. If the claim holds, the practical consequence is a single trained network that can take accelerated dynamic acquisitions and output motion estimates directly, which would support cardiac function assessment, motion tracking, and MR-guided radiotherapy.

What carries the argument

The load-bearing object is the full differentiable chain from k-space mask to deformation field. An adaptive sampling network produces a binary sampling mask through a straight-through estimator, a training trick that lets discrete sampling decisions pass gradients; the undersampled data then pass through the vSHARP reconstruction network, an unrolled ADMM optimizer that alternates learned denoising with data-consistency steps; and finally a 2D U-Net registration model outputs displacement fields that warp each reconstructed frame toward the reference. A combined loss $L = \alpha L_{\text{rec}} + \beta L_{\text{reg}}$, with the registration term measuring similarity between the warped reconstruction and the reference plus a smoothness penalty on the fields, is backpropagated through warping, reconstruction, and sampling selection, which is what makes the earlier stages optimize for motion estimation rather than image quality alone.

What would settle it

Apply the predicted deformation fields to the fully sampled ground-truth moving frames instead of the reconstructions and measure their similarity to the reference; if joint end-to-end training no longer beats decoupled training on that ground-truth metric, the paper's central claim that joint training improves motion estimation itself would be refuted.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central discovery is that coupling the three stages changes what each stage learns. The adaptive sampler is not merely maximizing reconstruction fidelity; it selects k-space lines that help the later registration align the moving frames to the reference. The reconstruction network is pushed to preserve the structures the registration needs, rather than only improving pixel-level image quality, and the registration module in turn provides the training signal that shapes both earlier stages. The paper reports that this joint optimization significantly improves registration quality compared with decoupled training, that phase-specific learned sampling beats unified and fixed patterns, that the unrolled vSHARP reconstruction supports registration better than the alternative VarNet, and that equal weighting of the reconstruction and registration losses gives the best balance between faithful images and accurate motion fields.

Load-bearing premise

The load-bearing premise is that the similarity between the warped reconstructed moving image and the reference image is a faithful measure of deformation-field accuracy, even though that score also improves whenever the reconstructed image itself resembles the reference.

Editorial extensions

If this is right

  • A single trained network can go directly from undersampled dynamic k-space to deformation fields, replacing separate acquisition-planning, reconstruction, and registration steps at inference.
  • Joint training makes the reconstruction subnetwork behave differently from a reconstruction-only network: it is shaped by what the registration needs, which is why decoupled training leaves registration quality on the table.
  • The choice of reconstruction backbone matters for motion estimation; the unrolled vSHARP design supports registration better than the tested VarNet configuration.
  • Phase-specific learned sampling patterns improve motion estimates over unified or fixed equispaced patterns, though equispaced initialization does not add a registration benefit.
  • The loss weights must be balanced: over-weighting registration can drive the reconstructor to imitate the reference image rather than the actual moving frame, inflating registration scores while corrupting the reconstruction.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The reported registration scores are computed on warped reconstructed images, so part of the apparent registration gain may be reconstruction gain; the authors acknowledge that registering fully sampled ground-truth moving images would be a more objective test.
  • If the modularity claim is right, the registration block could be swapped for a segmentation-supervised or transformer-based model without retraining the sampler, an experiment the paper does not run.
  • Because the adaptive sampler is trained with fully sampled k-space ground truth, prospective deployment on scans where full sampling is impossible would require a self-supervised reformulation of the sampling loss, a step the authors flag as future work.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 5 minor

Summary. This paper proposes an end-to-end deep learning framework for dynamic MRI that jointly optimizes adaptive k-space sampling, reconstruction, and deformable registration. The pipeline consists of a learned adaptive sampling mask generator, the vSHARP unrolled reconstruction network, and a 2D U-Net registration network, trained with a weighted combination of reconstruction and registration losses. Experiments on CMRxRecon 2023 cardiac cine and CMRxRecon 2024 aorta datasets compare registration modules, reconstruction backbones, sampling strategies, loss-weight configurations, and joint versus decoupled training. The evaluation metric is the similarity between the warped reconstructed moving image and the reference image (Eq. 18). The authors report that joint training and the proposed components improve registration quality and motion estimation, with the framework marketed as the first end-to-end integration of adaptive sampling, reconstruction, and registration.

Significance. If substantiated, the proposed pipeline would be a practically useful contribution to dynamic MRI, where undersampling and motion are central challenges. The modular design allowing interchangeable reconstruction and registration networks is attractive, and the paper provides a substantial set of ablations on both in-distribution and out-of-distribution data. The forward models and loss functions are clearly specified, and the authors are transparent about several limitations, including the reliance on fully sampled ground truth and the conflation of reconstruction and registration quality in the current evaluation. However, the central claim of improved motion estimation rests on an evaluation metric that is identical in form to the training loss, and the paper itself documents a failure mode in which the metric is maximized by reconstructing the reference image rather than by estimating correct deformations.

major comments (3)
  1. [Sec. 4.6 and Fig. 5] The registration-quality evaluation metric in Eq. 18 is the same similarity measure used as the registration training loss in Eq. 16 (SSIM+L1 between the warped reconstructed moving image and the reference). Consequently, the reported improvements in 'registration quality'—including the central claim in Sec. 4.7 that joint training significantly improves registration over decoupled training—do not necessarily reflect more accurate deformation fields. The pipeline can maximize the metric by driving the reconstructed moving image toward the reference, yielding near-perfect warped similarity with trivial motion. This is not hypothetical: the alpha=0, beta=1 experiment in Tables C1 and C2 and Figures C7/C8 shows registration SSIM rising to 0.987 while reconstruction SSIM collapses to 0.479, exactly the collapse mode described. The paper's own Sec. 5 statement that 'our results reflect both registration and reconstruction quality' and the suggestion to evaluate registered ground-truth moving images are acknowledgements that the main claim is currently untested. The revision should include an independent evaluation of the deformation fields—for example, warping the fully sampled ground-truth moving images with the predicted fields, or using synthetic/annotated data with known deformations—before the motion-estimation claims can be supported.
  2. [Sec. 5] The comparison of registration modules (proposed U-Net, VoxelMorph, TransMorph, optical flow, DEMONS) is affected by the same confound. Because Eq. 18 is computed on the registered reconstructed images, differences between registration modules can arise from different interactions with the jointly trained reconstruction module rather than from the modules' intrinsic motion-estimation capability. The conclusion that the proposed lightweight model yields 'more accurate motion estimation' than the baselines is therefore not supported by the presented evidence, even if the numbers are correct. The independent metric proposed above should be reported for all registration baselines (or at least for the learned ones) to establish the claimed advantage.
  3. [Sec. 4.7] The paper explicitly concedes that 'our results reflect both registration and reconstruction quality' and that 'evaluating motion estimation independently by comparing registered ground truth moving images... would provide a more objective assessment.' This limitation is load-bearing, not a minor caveat; it undermines the abstract's and conclusion's statements that the framework enables 'accurate deformation field estimation' and 'robust motion estimation from undersampled dynamic data.' Given the authors themselves identify the remedy, the revision should implement that evaluation rather than defer it to future work. Until then, the central claim is not directly validated.
minor comments (5)
  1. [Sec. 4.4, Eq. 18] The symbol m is used both for the averaged metric in Eq. 18 and for the mask operator M in Eq. 2; renaming one would avoid confusion.
  2. [Sec. 3.3, Eq. 17] The evaluation metric does not include the smoothness term Lsmooth, so the reported numbers do not capture whether the estimated deformations are physically plausible; reporting a regularity metric (e.g., percentage of folding or mean Jacobian determinant) would strengthen the motion-quality assessment.
  3. [Appendix C, Figs. C1-C10] Qualitative results show warped images but never visualize the estimated deformation fields; showing vector or grid overlays of the predicted motion would help the reader assess the behavior of the registration component.
  4. [Sec. 4.6] The comparison with VoxelMorph and TransMorph does not include information about the number of trainable parameters or training cost, which would be useful for judging the 'lightweight' claim.
  5. [Appendix C, Table C5] The heading 'Registration uantitative results' contains a typo ('uantitative' should be 'quantitative').

Circularity Check

1 steps flagged · score 5.0 of 10

Registration-quality metric (Eq. 18) is the same similarity objective optimized in Lreg (Eq. 16), so joint-training registration gains are partly self-fulfilling.

  1. fitted input called prediction [Section 4.4 (Evaluation, Eq. 18); cf. Section 3.3 (Eq. 16); acknowledged in Section 5]
    "We assess estimated motion quality by evaluating the similarity of the registered moving images (warped reconstructed images using predicted deformation fields) to the reference image."

    The evaluation metric in Eq. 18 compares W(|xhat_mov|, phi) with xref, which is exactly the Lsim term in the training loss Lreg of Eq. 16 (plus smoothness). Because the whole pipeline, including the reconstruction network, is trained end-to-end to minimize Lreg, the reported 'registration quality' improvement from joint training is an improvement in the training objective itself. The paper's alpha=0 ablation confirms the metric is gameable: registration SSIM rises to 0.987 while reconstruction SSIM collapses to 0.479 (Tables C1/C2), meaning the model can maximize Eq. 18 by making xhat_mov approximate xref under a trivial deformation field.

full rationale

One genuine but partial circularity affects the central registration-quality claim. Eq. 18 defines motion quality as the similarity of the warped reconstructed moving image to the reference, which is the same similarity term optimized as Lreg in Eq. 16. Because the reconstruction module is jointly trained, the pipeline can improve Eq. 18 either by genuinely aligning motion or by making the reconstruction resemble the reference; the alpha=0 ablation documents the latter failure mode. This makes the joint-versus-decoupled registration comparison and the adaptive-sampling registration comparisons partly self-fulfilling. However, the circularity is not total: reconstruction quality is evaluated independently against ground-truth moving images (Tables C2, C4, C6, C8), the default alpha=beta=1 operating point retains substantial reconstruction fidelity, and the component-wise engineering comparisons (vSHARP vs. VarNet, learned vs. fixed sampling, phase-specific vs. unified) rest on additional implementation choices and external baselines. The self-citations to [60] and [57] are used as modular components and are not load-bearing circular premises; they are compared against independent methods. The score reflects the one metric-level reduction, not a collapse of the entire derivation chain.

Assumptions & free parameters 4 free parameters · 4 assumptions · 0 invented entities

The paper introduces no new physical entities. The main hand-chosen inputs are the loss weights, ACS fraction, and unrolled reconstruction iterations; the key domain assumptions concern the validity of retrospective undersampling and the similarity metric as a proxy for motion accuracy.

free parameters (4)
  • Loss weights alpha and beta = alpha=1, beta=1 (default; explored alpha in {0, 1, 1.5, 2} and beta in {1, 2, 3})
    Hand-chosen to balance reconstruction and registration terms in Eq. 14; the paper finds balanced weighting yields the best motion estimates.
  • ACS fraction = 4% of k-space lines
    Fixed central low-frequency region retained for sensitivity estimation and to initialize the adaptive sampler; not ablated in the paper.
  • Number of ADS cascades = 1 (versus 2 in prior work [60])
    Reduced from the E2E-ADS-Recon configuration; no ablation is performed on this choice.
  • vSHARP iterations (T, Tx) = T=10, Tx=6
    Unrolled optimization steps and data-consistency steps taken from vSHARP [57]; not ablated in this paper.
assumptions (4)
  • domain assumption Retrospectively undersampled fully-sampled k-space is a valid proxy for prospectively undersampled acquisition.
    All training and evaluation use retrospective masks on fully sampled data; Sec. 5 lists this as a limitation.
  • domain assumption The diffeomorphic deformation model (scaling and squaring plus diffusion smoothness) can accurately represent cardiac motion.
    Registration uses integrated displacement fields with Lsmooth (Eq. 17), assuming smooth, invertible motion.
  • domain assumption The 6th cardiac phase is a consistent end-systolic reference across cases.
    Sec. 4.3 selects frame 6 as reference because it can be reliably ECG-triggered, providing a stable anatomical structure.
  • ad hoc to paper Similarity of warped images to the reference measures clinically meaningful motion estimation.
    The primary evaluation metric (Eq. 18) equals the training loss; the paper acknowledges this conflation in Sec. 5 and suggests more objective assessments for future work.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Deep End-to-end Adaptive k-Space Sampling, Reconstruction, and Registration for Dynamic MRI." pith.science (2026). https://pith.science/paper/5GFBRZIW

@misc{pith2026241118249,
  author       = {Pith},
  title        = {Pith review of: Deep End-to-end Adaptive k-Space Sampling, Reconstruction, and Registration for Dynamic MRI},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/5GFBRZIW}},
  note         = {Machine review of arXiv:2411.18249}
}
read the original abstract

Dynamic MRI enables a range of clinical applications, including cardiac function assessment, organ motion tracking, and radiotherapy guidance. However, fully sampling the dynamic k-space data is often infeasible due to time constraints and physiological motion such as respiratory and cardiac motion. This necessitates undersampling, which degrades the quality of reconstructed images. Poor image quality not only hinders visualization but also impairs the estimation of deformation fields, crucial for registering dynamic (moving) images to a static reference image. This registration enables tasks such as motion correction, treatment planning, and quantitative analysis in applications like cardiac imaging and MR-guided radiotherapy. To overcome the challenges posed by undersampling and motion, we introduce an end-to-end deep learning (DL) framework that integrates adaptive dynamic k-space sampling, reconstruction, and registration. Our approach begins with a DL-based adaptive sampling strategy, optimizing dynamic k-space acquisition to capture the most relevant data for each specific case. This is followed by a DL-based reconstruction module that produces images optimized for accurate deformation field estimation from the undersampled moving data. Finally, a registration module estimates the deformation fields aligning the reconstructed dynamic images with a static reference. The proposed framework is independent of specific reconstruction and registration modules allowing for plug-and-play integration of these components. The entire framework is jointly trained using a combination of supervised and unsupervised loss functions, enabling end-to-end optimization for improved performance across all components. Through controlled experiments and ablation studies, we validate each component, demonstrating that each choice contributes to robust motion estimation from undersampled dynamic data.

Figures

Figures reproduced from arXiv: 2411.18249 by the authors.

Figure 1
Figure 1. Overview of our proposed pipeline: Utilizing initial dynamic (moving) [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Adaptive Sampling Model (Aω) framework. Sensitivity profiles from the sensitivity model (Sσ) and initial undersampled moving data (y˜M0 mov) are pro￾cessed by a U-Net-like encoder followed by a MLP, generating sampling proba￾bilities. These are rescaled for the specified acceleration R and binarized via a straight-through estimator creating an adapted binary dynamic sampling mask. Here we assume y˜M0 mov = y˜Macs mo… view at source ↗
Figure 3
Figure 3. A 2D U-Net plays the role of the registration model. It takes as input [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗
Figures from the paper (8 more)
Figure 4
Figure 4. Figure 4: End-to-end pipeline of the proposed method. The process starts with coil [PITH_FULL_IMAGE:figures/full_fig_p009_4.png]
Figure 5
Figure 5. Figure 5: Comparison of registration performance across different methods. Top: [PITH_FULL_IMAGE:figures/full_fig_p013_5.png]
Figure 6
Figure 6. Figure 6: Impact of reconstruction model on registration results. Top: Cine. Bottom: [PITH_FULL_IMAGE:figures/full_fig_p014_6.png]
Figure 7
Figure 7. Figure 7: Impact of phase-specific vs unified adaptive sampling with ACS and equi [PITH_FULL_IMAGE:figures/full_fig_p014_7.png]
Figure 8
Figure 8. Figure 8: Impact of learned vs fixed non-adaptive sampling schemes on registration [PITH_FULL_IMAGE:figures/full_fig_p014_8.png]
Figure 9
Figure 9. Figure 9: Impact of joint (J) vs decoupled (D) training, and varying loss weights [PITH_FULL_IMAGE:figures/full_fig_p015_9.png]
Figure 10
Figure 10. Figure 10: Impact of loss choice (cine dataset). Top: Reconstruction performance. [PITH_FULL_IMAGE:figures/full_fig_p015_10.png]
Figure 11
Figure 11. Figure 11: Example results for two cases, shown at various temporal frames ( [PITH_FULL_IMAGE:figures/full_fig_p016_11.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

74 extracted references · 56 canonical work pages

  1. [1]

    IEEE transactions on medical imaging 37(6), 1322–1332 (2018)

    Adler, J., ¨Oktem, O.: Learned primal-dual reconstruction. IEEE transactions on medical imaging 37(6), 1322–1332 (2018)

  2. [2]

    IEEE transactions on medical imaging 38(2), 394–405 (2018)

    Aggarwal, H.K., Mani, M.P., Jacob, M.: Modl: Model-based deep learning architec- ture for inverse problems. IEEE transactions on medical imaging 38(2), 394–405 (2018)

  3. [3]

    Bahadir, C.D., Dalca, A.V., Sabuncu, M.R.: Learning-Based Optimization of the Under-Sampling Pattern in MRI, p. 780–792. Springer International Publishing (2019). https://doi.org/10.1007/978-3-030-20351-1_61

  4. [4]

    In: Medical Imaging with Deep Learning (2022)

    Bakker, T., Muckley, M.J., Romero-Soriano, A., Drozdzal, M., Pineda, L.: On learning adaptive acquisition policies for undersampled multi-coil MRI reconstruc- tion. In: Medical Imaging with Deep Learning (2022)

  5. [5]

    IEEE transactions on medical imaging 38(8), 1788–1800 (2019)

    Balakrishnan, G., Zhao, A., Sabuncu, M.R., Guttag, J., Dalca, A.V.: Voxelmorph: a learning framework for deformable medical image registration. IEEE transactions on medical imaging 38(8), 1788–1800 (2019)

  6. [6]

    https://doi.org/10.3389/fnins.2022

    Beauferris, Y., Teuwen, J., Karkalousos, D., Moriakov, N., Caan, M., Yi- asemis, G., Rodrigues, L., Lopes, A., Pedrini, H., Rittner, L., Dan- necker, M., Studenyak, V., Gr¨ oger, F., Vyas, D., Faghih-Roohi, S., Ku- mar Jethi, A., Chandra Raju, J., Sivaprakasam, M., Lasby, M., Nogovit- syn, N., Loos, W., Frayne, R., Souza, R.: Multi-coil mri reconstruction...

  7. [7]

    arXiv preprint arXiv:1308.3432 (2013)

    Bengio, Y., L´ eonard, N., Courville, A.: Estimating or propagating gradi- ents through stochastic neurons for conditional computation. arXiv preprint arXiv:1308.3432 (2013)

  8. [8]

    Physics and Imaging in Radiation Oncology 24, 167–172 (2022)

    Bertelsen, A., Bernchou, U., Schytte, T., Brink, C., Mahmood, F.: The effect of respiration-induced target motion on 3d magnetic resonance images used to guide radiotherapy. Physics and Imaging in Radiation Oncology 24, 167–172 (2022)

Show all 74 references
  1. [9]

    Medical image analysis 82, 102615 (2022)

    Chen, J., Frey, E.C., He, Y., Segars, W.P., Li, Y., Du, Y.: Transmorph: Transformer for unsupervised medical image registration. Medical image analysis 82, 102615 (2022)

  2. [10]

    Progress in Biomedical Engineering 3(1), 012003 (2021)

    Chen, X., Diaz-Pinto, A., Ravikumar, N., Frangi, A.F.: Deep learning in medical image registration. Progress in Biomedical Engineering 3(1), 012003 (2021)

  3. [11]

    Proceedings of the IEEE 110(2), 224–245 (2022)

    Chen, Y., Sch¨ onlieb, C.B., Li` o, P., Leiner, T., Dragotti, P.L., Wang, G., Rueckert, D., Firmin, D., Yang, G.: Ai-based reconstruction for fast mri—a systematic review and meta-analysis. Proceedings of the IEEE 110(2), 224–245 (2022)

  4. [12]

    Medical Image Analysis 12(4), 427–441 (2008) 20 G

    Commowick, O., Arsigny, V., Isambert, A., Costa, J., Dhermain, F., Bidault, F., Bondiau, P.Y., Ayache, N., Malandain, G.: An efficient locally affine framework for the smooth registration of anatomical structures. Medical Image Analysis 12(4), 427–441 (2008) 20 G. Yiasemis et al

  5. [13]

    Medical Image Analysis 68, 101941 (2021)

    Corona, V., Aviles-Rivero, A., Debroux, N., Le Guyader, C., Sch¨ onlieb, C.B.: Vari- ational multi-task mri reconstruction: Joint reconstruction, registration and super- resolution. Medical Image Analysis 68, 101941 (2021)

  6. [14]

    Magnetic resonance in medicine 80(5), 2188–2201 (2018)

    Eo, T., Jun, Y., Kim, T., Jang, J., Lee, H.J., Hwang, D.: Kiki-net: cross-domain convolutional neural networks for reconstructing undersampled magnetic resonance images. Magnetic resonance in medicine 80(5), 2188–2201 (2018)

  7. [15]

    Mag- netic Resonance in Medicine: An Official Journal of the International Society for Magnetic Resonance in Medicine 59(2), 365–373 (2008)

    Gamper, U., Boesiger, P., Kozerke, S.: Compressed sensing in dynamic mri. Mag- netic Resonance in Medicine: An Official Journal of the International Society for Magnetic Resonance in Medicine 59(2), 365–373 (2008)

  8. [16]

    Gautam, S., Li, A., Ravishankar, S.: Patient-adaptive and learned mri data under- sampling using neighborhood clustering (2024)

  9. [17]

    Magnetic Resonance in Medicine 79(6), 3055–3071 (nov 2017)

    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 Medicine 79(6), 3055–3071 (nov 2017). https:// doi.org/10.1002/mrm.26977, https://doi.org...

  10. [18]

    Hering, A., Kuckertz, S., Heldmann, S., Heinrich, M.: Enhancing label-driven deep deformable image registration with local distance metrics for state-of-the-art car- diac motion tracking (2018), https://arxiv.org/abs/1812.01859

  11. [19]

    Clinical Oncology 30(11), 711–719 (2018)

    Hunt, A., Hansen, V., Oelfke, U., Nill, S., Hafeez, S.: Adaptive radiotherapy enabled by mri guidance. Clinical Oncology 30(11), 711–719 (2018). https://doi.org/ https://doi.org/10.1016/j.clon.2018.08.001, https://www.sciencedirect. com/science/article/pii/S0936655518304084, m...

  12. [20]

    Fron- tiers in cardiovascular medicine 9, 826283 (2022)

    Ismail, T.F., Strugnell, W., Coletti, C., Boˇ zi´ c-Iven, M., Weingaertner, S., Ham- mernik, K., Correia, T., Kuestner, T.: Cardiac mr: from theory to practice. Fron- tiers in cardiovascular medicine 9, 826283 (2022)

  13. [21]

    Advances in neural information processing systems 28 (2015)

    Jaderberg, M., Simonyan, K., Zisserman, A., et al.: Spatial transformer networks. Advances in neural information processing systems 28 (2015)

  14. [22]

    Cardiology Research and Practice 2018, 1–15 (aug 2018)

    Khalil, A., Ng, S.C., Liew, Y.M., Lai, K.W.: An overview on image registra- tion techniques for cardiac diagnosis and treatment. Cardiology Research and Practice 2018, 1–15 (aug 2018). https://doi.org/10.1155/2018/1437125, http: //dx.doi.org/10.1155/2018/1437125

  15. [23]

    Magnetic resonance in medicine 84(6), 3054–3070 (2020)

    Knoll, F., Murrell, T., Sriram, A., Yakubova, N., Zbontar, J., Rabbat, M., De- fazio, A., Muckley, M.J., Sodickson, D.K., Zitnick, C.L., et al.: Advancing machine learning for mr image reconstruction with an open competition: Overview of the 2019 fastmri challenge. Magnetic re...

  16. [24]

    Modern signal processing 46, 161–184 (2003)

    Kostelec, P.J., Periaswamy, S.: Image registration for mri. Modern signal processing 46, 161–184 (2003)

  17. [25]

    In: IEEE International Conference on Image Processing 2005

    Le Besnerais, G., Champagnat, F.: Dense optical flow by iterative local window registration. In: IEEE International Conference on Image Processing 2005. vol. 1, pp. I–137 (2005). https://doi.org/10.1109/ICIP.2005.1529706

  18. [26]

    In: Proceedings of the IEEE/CVF international conference on computer vision

    Liu, Z., Lin, Y., Cao, Y., Hu, H., Wei, Y., Zhang, Z., Lin, S., Guo, B.: Swin transformer: Hierarchical vision transformer using shifted windows. In: Proceedings of the IEEE/CVF international conference on computer vision. pp. 10012–10022 (2021)

  19. [27]

    Computerized Med- ical Imaging and Graphics 113, 102348 (2024)

    Lønning, K., Caan, M.W., Nowee, M.E., Sonke, J.J.: Dynamic recurrent inference machines for accelerated mri-guided radiotherapy of the liver. Computerized Med- ical Imaging and Graphics 113, 102348 (2024)

  20. [28]

    Frontiers in Neuroinformatics 7 (2013)

    Lowekamp, B.C., Chen, D.T., Ibanez, L., Blezek, D.: The design of simpleitk. Frontiers in Neuroinformatics 7 (2013). https://doi.org/10.3389/fninf.2013. 00045, https://www.frontiersin.org/journals/neuroinformatics/articles/ 10.3389/fninf.2013.00045 E2E Adaptive Sampling, Recon...

  21. [29]

    IEEE Signal Processing Magazine 25(2), 72–82 (2008)

    Lustig, M., Donoho, D.L., Santos, J.M., Pauly, J.M.: Compressed sensing mri. IEEE Signal Processing Magazine 25(2), 72–82 (2008). https://doi.org/10. 1109/MSP.2007.914728

  22. [30]

    Medical Image Analysis p

    Lyu, J., Qin, C., Wang, S., Wang, F., Li, Y., Wang, Z., Guo, K., Ouyang, C., T¨ anzer, M., Liu, M., et al.: The state-of-the-art in cardiac mri reconstruction: Results of the cmrxrecon challenge in miccai 2023. Medical Image Analysis p. 103485 (2025)

  23. [31]

    JACC: Cardiovascular Imaging 8(2), 148–157 (2015)

    Mada, R.O., Lysyansky, P., Daraban, A.M., Duchenne, J., Voigt, J.U.: How to define end-diastole and end-systole?: Impact of timing on strain measure- ments. JACC: Cardiovascular Imaging 8(2), 148–157 (2015). https://doi.org/ https://doi.org/10.1016/j.jcmg.2014.10.010, https://...

  24. [32]

    Bio- engineering 10(3) (2023)

    Oscanoa, J.A., Middione, M.J., Alkan, C., Yurt, M., Loecher, M., Vasanawala, S.S., Ennis, D.B.: Deep learning-based reconstruction for cardiac mri: A review. Bio- engineering 10(3) (2023). https://doi.org/10.3390/bioengineering10030334, https://www.mdpi.com/2306-5354/10/3/334

  25. [33]

    Pal, A., Rathi, Y.: A review and experimental evaluation of deep learning methods for MRI reconstruction (mar 2022)

  26. [34]

    Paszke, A., Gross, S., Massa, F., Lerer, A., Bradbury, J., Chanan, G., Killeen, T., Lin, Z., Gimelshein, N., Antiga, L., et al.: Pytorch: An imperative style, high- performance deep learning library (2019)

  27. [35]

    In: Lecture Notes in Computer Science

    Ronneberger, O., Fischer, P., Brox, T.: U-net: Convolutional networks for biomedical image segmentation. In: Lecture Notes in Computer Science. pp. 234–241. Springer International Publishing (2015). https://doi.org/10.1007/ 978-3-319-24574-4_28 , https://doi.org/10.1007/978-3-...

  28. [36]

    Physics in Medicine & Biology 59(3), 733 (2014)

    Saleh, Z.H., Apte, A.P., Sharp, G.C., Shusharina, N.P., Wang, Y., Veeraraghavan, H., Thor, M., Muren, L.P., Rao, S.S., Lee, N.Y., et al.: The distance discordance metric—a novel approach to quantifying spatial uncertainties in intra-and inter- patient deformable image registra...

  29. [37]

    Proceedings of the IRE 37(1), 10–21 (1949)

    Shannon, C.: Communication in the presence of noise. Proceedings of the IRE 37(1), 10–21 (1949). https://doi.org/10.1109/JRPROC.1949.232969

  30. [38]

    In: Medical Imaging with Deep Learning (2023)

    Shor, T.: Multi PILOT: Feasible learned multiple acquisition trajectories for dy- namic MRI. In: Medical Imaging with Deep Learning (2023)

  31. [39]

    In: Lecture Notes in Computer Science

    Sriram, A., Zbontar, J., Murrell, T., et al.: End-to-end variational networks for accelerated mri reconstruction. In: Lecture Notes in Computer Science. p. 64–73. Springer International Publishing (2020). https://doi.org/10.1007/ 978-3-030-59713-9_7 , http://dx.doi.org/10.1007...

  32. [40]

    Medical Image Analysis 97, 103249 (2024)

    Sun, S., Han, K., You, C., Tang, H., Kong, D., Naushad, J., Yan, X., Ma, H., Khosravi, P., Duncan, J.S., et al.: Medical image registration via neural fields. Medical Image Analysis 97, 103249 (2024)

  33. [41]

    Medical physics 48(11), 6597–6613 (2021)

    Terpstra, M.L., Maspero, M., Bruijnen, T., Verhoeff, J.J., Lagendijk, J.J., van den Berg, C.A.: Real-time 3d motion estimation from undersampled mri using multi- resolution neural networks. Medical physics 48(11), 6597–6613 (2021)

  34. [42]

    Medical Image Analysis 80, 102509 (2022)

    Terpstra, M.L., Maspero, M., Sbrizzi, A., van den Berg, C.A.: ⊥-loss: A symmetric loss function for magnetic resonance imaging reconstruction and image registration with deep learning. Medical Image Analysis 80, 102509 (2022). https://doi.org/https://doi.org/10.1016/j.media.20...

  35. [43]

    Medical image analysis 2(3), 243–260 (1998) 22 G

    Thirion, J.P.: Image matching as a diffusion process: an analogy with maxwell’s demons. Medical image analysis 2(3), 243–260 (1998) 22 G. Yiasemis et al

  36. [44]

    Magnetic Resonance in Medicine: An Official Journal of the International Society for Magnetic Resonance in Medicine 50(5), 1031–1042 (2003)

    Tsao, J., Boesiger, P., Pruessmann, K.P.: k-t blast and k-t sense: dynamic mri with high frame rate exploiting spatiotemporal correlations. Magnetic Resonance in Medicine: An Official Journal of the International Society for Magnetic Resonance in Medicine 50(5), 1031–1042 (2003)

  37. [45]

    In: XVth international conference on the use of computers in radiation therapy (ICCR)

    Vandemeulebroucke, J., Sarrut, D., Clarysse, P., et al.: The popi-model, a point- validated pixel-based breathing thorax model. In: XVth international conference on the use of computers in radiation therapy (ICCR). vol. 2, pp. 195–199 (2007)

  38. [46]

    arXiv preprint arXiv:2202.05267 (2022)

    Waddington, D.E., Hindley, N., Koonjoo, N., Chiu, C., Reynolds, T., Liu, P.Z., Zhu, B., Bhutto, D., Paganelli, C., Keall, P.J., et al.: On real-time image reconstruction with neural networks for mri-guided radiotherapy. arXiv preprint arXiv:2202.05267 (2022)

  39. [47]

    PeerJ 2, e453 (2014)

    Van der Walt, S., Sch¨ onberger, J.L., Nunez-Iglesias, J., Boulogne, F., Warner, J.D., Yager, N., Gouillart, E., Yu, T.: scikit-image: image processing in python. PeerJ 2, e453 (2014)

  40. [48]

    Phenomics 1(4), 151–170 (jul 2021)

    Wang, C., Li, Y., Lv, J., Jin, J., Hu, X., Kuang, X., Chen, W., Wang, H.: Rec- ommendation for cardiac magnetic resonance imaging-based phenotypic study: Imaging part. Phenomics 1(4), 151–170 (jul 2021). https://doi.org/10.1007/ s43657-021-00018-x

  41. [49]

    https://doi.org/10.48550/ARXIV

    Wang, C., et al.: Cmrxrecon: An open cardiac mri dataset for the competition of accelerated image reconstruction (2023). https://doi.org/10.48550/ARXIV. 2309.10836

  42. [50]

    org/abs/2503.03971

    Wang, F., Wang, Z., Li, Y., Lyu, J., Qin, C., Wang, S., Guo, K., Sun, M., Huang, M., Zhang, H., T¨ anzer, M., Li, Q., Chen, X., Huang, J., Wu, Y., Lyu, Y., Sun, L., Li, Q., Xu, Z., Xin, B., Metaxas, D.N., Hamdani, K.A., Nabavi, S., Yiasemis, G., Teuwen, J., Zhang, Z., Wang, S....

  43. [51]

    Wang, Z., Wang, F., Qin, C., Lyu, J., Cheng, O., Wang, S., Li, Y., Yu, M., Zhang, H., Guo, K., Shi, Z., Li, Q., Xu, Z., Zhang, Y., Li, H., Hua, S., Chen, B., Sun, L., Sun, M., Li, Q., Chu, Y.H., Bai, W., Qin, J., Zhuang, X., Prieto, C., Young, A., Markl, M., Wang, H., Wu, L., ...

  44. [52]

    In: 2020 IEEE 17th International Symposium on Biomedical Imaging (ISBI)

    Yaman, B., Hosseini, S.A.H., Moeller, S., Ellermann, J., U˘ gurbil, K., Ak¸ cakaya, M.: Self-supervised physics-based deep learning mri reconstruction without fully- sampled data. In: 2020 IEEE 17th International Symposium on Biomedical Imaging (ISBI). pp. 921–925 (2020). http...

  45. [53]

    Computers in Bi- ology and Medicine 167, 107619 (2023)

    Yan, Y., Yang, T., Zhao, X., Jiao, C., Yang, A., Miao, J.: Dc-siamnet: Deep con- trastive siamese network for self-supervised mri reconstruction. Computers in Bi- ology and Medicine 167, 107619 (2023)

  46. [54]

    IEEE Transactions on Biomedical Engineering (2024)

    Yang, F., Xue, Z., Lu, H., Xu, J., Chen, H., Chen, Z., Emu, Y., Aburas, A., Gao, J., Gao, C., et al.: Robust fast inter-bin image registration for undersampled coronary mri based on a learned motion prior. IEEE Transactions on Biomedical Engineering (2024)

  47. [55]

    Yiasemis, G., Moriakov, N., S´ anchez, C.I., Sonke, J.J., Teuwen, J.: Jssl: Joint supervised and self-supervised learning for mri reconstruction (2023) E2E Adaptive Sampling, Reconstruction, and Registration for Dynamic MRI 23

  48. [56]

    arXiv preprint arXiv:2411.01291 (2024)

    Yiasemis, G., Moriakov, N., Sonke, J.J., Teuwen, J.: Deep multi-contrast cardiac mri reconstruction via vsharp with auxiliary refinement network. arXiv preprint arXiv:2411.01291 (2024)

  49. [57]

    Magnetic Resonance Imaging 115, 110266 (2025)

    Yiasemis, G., Moriakov, N., Sonke, J.J., Teuwen, J.: vsharp: Variable splitting half-quadratic admm algorithm for reconstruction of inverse-problems. Magnetic Resonance Imaging 115, 110266 (2025). https://doi.org/https://doi.org/10. 1016/j.mri.2024.110266, https://www.scienced...

  50. [58]

    Magnetic Resonance Imaging 107, 33–46 (2024)

    Yiasemis, G., S´ anchez, C.I., Sonke, J.J., Teuwen, J.: On retrospective k-space sub- sampling schemes for deep mri reconstruction. Magnetic Resonance Imaging 107, 33–46 (2024)

  51. [59]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)

    Yiasemis, G., Sonke, J.J., S´ anchez, C., Teuwen, J.: Recurrent variational network: A deep learning inverse problem solver applied to the task of accelerated mri re- construction. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). pp...

  52. [60]

    Yiasemis, G., Sonke, J.J., Teuwen, J.: End-to-end adaptive dynamic subsampling and reconstruction for cardiac mri (2024), https://arxiv.org/abs/2403.10346

  53. [61]

    arXiv preprint arXiv:2105.06460 (2021)

    Yin, T., Wu, Z., Sun, H., Dalca, A.V., Yue, Y., Bouman, K.L.: End-to-end sequen- tial sampling and reconstruction for mri. arXiv preprint arXiv:2105.06460 (2021)

  54. [62]

    IEEE Transactions on Medical Imaging 40(12), 3337–3348 (2021)

    Yoo, J., Jin, K.H., Gupta, H., Yerly, J., Stuber, M., Unser, M.: Time-dependent deep image prior for dynamic mri. IEEE Transactions on Medical Imaging 40(12), 3337–3348 (2021)

  55. [63]

    In: Hamprecht, F.A., Schn¨ orr, C., J¨ ahne, B

    Zach, C., Pock, T., Bischof, H.: A duality based approach for realtime tv-l1 optical flow. In: Hamprecht, F.A., Schn¨ orr, C., J¨ ahne, B. (eds.) Pattern Recognition. pp. 214–223. Springer Berlin Heidelberg, Berlin, Heidelberg (2007)

  56. [64]

    Zhang, J., Zhang, H., Wang, A., Zhang, Q., Sabuncu, M., Spincemaille, P., Nguyen, T.D., Wang, Y.: Extending loupe for k-space under-sampling pattern optimization in multi-coil mri (2020)

  57. [65]

    In: 2018 IEEE/ACM 26th international symposium on quality of service (IWQoS)

    Zhang, Z.: Improved adam optimizer for deep neural networks. In: 2018 IEEE/ACM 26th international symposium on quality of service (IWQoS). pp. 1–2. Ieee (2018) 24 G. Yiasemis et al. Supplementary Material A Methods - Additional Information A.1 Sensitivity Profile Network As de...

  58. [66]

    Autocalibration Signal (ACS): The central k-space region was retained, cov- ering 4% of the fully sampled data

  59. [67]

    ACS measurements (4% of the fully sampled data) were included in this sampling pattern

    Equispaced Sampling: Applied with a target acceleration factor of ( R − 4)×, where R is the target acceleration. ACS measurements (4% of the fully sampled data) were included in this sampling pattern. For adaptive sampling, the initial undersampled data informed the acquisitio...

  60. [68]

    Separate patterns were used for phase-specific experiments, while a unified pattern was applied across all phases in unified experiments

    Equispaced: Lines were evenly spaced to achieve the target acceleration, with the central offset randomized during training and fixed during inference. Separate patterns were used for phase-specific experiments, while a unified pattern was applied across all phases in unified ...

  61. [69]

    Each time-step represents a scheme per each phase

    kt-Equispaced: Used exclusively in phase-specific experiments, this strategy followed the method in [44], where equispaced sampling was performed per E2E Adaptive Sampling, Reconstruction, and Registration for Dynamic MRI 27 (a) Phase-specific equis- paced sampling pattern. Ea...

  62. [70]

    For each moving (dynamic) image, we performed a forward pass for each temporal phase along with the reference image, consistent with the original implementation in [5].image

    Voxelmorph with a 2D U-Net architecture comprising 4 scales with filter sizes of 16, 32, 64, and 128 was employed. For each moving (dynamic) image, we performed a forward pass for each temporal phase along with the reference image, consistent with the original implementation i...

  63. [71]

    The setup also includes Global Position-Sensitive Attention (GPSA) intervals and locality strength to focus on spatial regions

    Transmorph with a vision transformer model (ViT) inspired by the Swin Transformer [26], configured with a patch size of 10 × 10, an embedding dimension of 64, 8 layers, and 9 attention heads. The setup also includes Global Position-Sensitive Attention (GPSA) intervals and loca...

  64. [72]

    Iterative Lucas-Kanade (ILK) optical flow method from scikit-image, con- figured with a radius of 5, 3 warp iterations, prefiltering enabled, and without Gaussian smoothing. All other parameters are based on the default settings of the scikit-image in https://scikit-image.org/...

  65. [73]

    All other parameters are based on the default settings of the scikit-image in https://scikit-image.org/ docs/stable/api/skimage.registration.html#skimage.registration

    Total Variation L1 optical flow method from scikit-image, configured with an attachment weight of 15, tightness parameter of 0.3, 3 warp iterations, 5 main iterations, and a tolerance of 1×10−2. All other parameters are based on the default settings of the scikit-image in http...

  66. [74]

    See https://simpleitk.org/doxygen/latest/html/classitk_1_1simple_ 1_1DemonsRegistrationFilter.html

    Demons registration method configured with 10 iterations, a Gaussian smooth- ing standard deviation of 1.0, and smoothing applied to the displacement field, following the standard settings of the SimpleITK Demons algorithm. See https://simpleitk.org/doxygen/latest/html/classit...

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.