Pith. sign in

REVIEW 4 major objections 3 minor 45 references

EMORe: Motion-Robust 5D MRI Reconstruction via Expectation-Maximization-Guided Binning Correction and Outlier Rejection

T0 review · 4 major / 3 minor · reviewed 2026-08-06 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read EMORe: a reconstruction method that corrects motion-bin assignments and rejects outlier readouts in self-gated 5D cardiac MRI.

desk verdict EMORe is a well-validated EM-based binning correction for 5D cardiac MRI that deserves peer review, but its robustness claim needs a sensitivity analysis on the binning prior. read the letter →

arxiv 2507.23224 v1 pith:OXMCDHWH submitted 2025-07-31 eess.IV eess.SP

Syed M. Arshad , Lee C. Potter , Yingmin Liu , Christopher Crabtree , Matthew S. Tong , Rizwan Ahmad This is my paper · ORCID
classification eess.IVeess.SP
keywords 5DcardiacMRIself-gatingexpectation-maximizationbinningcorrectionoutlierrejectioncompressedsensingmotionartifactsfree-breathing
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

This paper proposes EMORe, a reconstruction method for free-running, free-breathing self-gated 5D cardiac MRI that treats imperfect retrospective motion binning as a correctable error rather than a fixed input. The authors aim to show that by iteratively re-estimating each k-space readout's probability of belonging to each of 80 cardiorespiratory motion bins, plus an explicit outlier bin for motion-corrupted data, while simultaneously improving the image estimate, residual blur and bulk-motion artifacts can be substantially reduced. If true, this matters because 5D cardiac MRI currently relies on self-gating binning that is corrupted by inaccurate signal extraction, irregular breathing, and sporadic patient motion such as coughing, and these errors limit diagnostic confidence. The paper validates the claim in simulated MRXCAT phantom studies and in 13 in vivo scans, reporting statistically significant gains over compressed sensing in image sharpness, structural similarity, and blinded artifact scores. The method is a drop-in reconstruction upgrade at a modest increase in computation time.

What carries the argument

The central mechanism is an expectation-maximization loop in which the true motion-bin assignment of every readout is a latent variable. In the E-step, the posterior probability that readout $n$ belongs to bin $k$ is computed by Bayes' rule from a Gaussian k-space residual likelihood and an informative self-gating prior $\theta(n,k)$; the $(K+1)$-th 'bin' is an outlier state whose likelihood is a constant $\exp(-\tau^2/\sigma^2)$ with $\tau=3\sigma$. In the M-step, the images for all $K$ motion states are updated by minimizing a weighted least-squares data-fidelity term plus anisotropic total-variation regularization along spatial, cardiac, and respiratory dimensions, solved approximately with ADMM. The binning prior $\theta(n,k)$ assigns probability $0.85$ to the initial self-gating assignment, $0.05$ to the outlier bin, and spreads the remainder over the other bins, which stabilizes the high-dimensional EM loop but caps how far a readout can move from its original bin.

What would settle it

Run EMORe on a phantom dataset in which a controlled fraction of readouts (e.g., 20–40%) is deliberately assigned to the wrong motion bin by corrupting the self-gating signal, and compare PSNR and Brier score against compressed sensing; the central claim would be refuted if EMORe's advantage disappears or if its final bin assignments remain as wrong as the initialization. A minimal version is to lower the prior weight $\alpha_g$ from $0.85$ to $0.5$ and show that EMORe's output quality collapses once the true bins are no longer protected.

Watch

Extended reading notes

Core claim

The paper's central claim is that residual motion artifacts in self-gated 5D cardiac MRI arise from two recoverable causes: valid readouts assigned to the wrong motion bin, and readouts corrupted by bulk motion that belong to no valid bin. By treating each readout's true bin assignment as a latent variable, EMORe uses an expectation-maximization loop that alternates between computing soft posterior probabilities of bin membership and updating the reconstructed images. The outlier bin, the (K+1)-th state, collects readouts inconsistent with all valid motion states, so corrupted data are rejected rather than blurring the image. In the MRXCAT phantom study, which includes 50 simulated scans spanning 0% to 70% bulk-motion corruption, EMORe outperforms standard compressed sensing in peak signal-to-noise ratio, structural similarity index, edge sharpness, and bin-assignment accuracy as measured by Brier score. In 13 in vivo scans, including three with instructed coughing, EMORe significantly improves blood-myocardium edge sharpness and expert artifact scores relative to compressed sensing.

Load-bearing premise

The method's correction power rests on the initial self-gating binning being correct for the large majority of readouts, because the prior pins $0.85$ of each readout's prior probability to its original bin; if systematic self-gating errors exceed roughly $15\%$, the EM loop is biased toward the wrong bins and may not recover them, a sensitivity the authors themselves note.

Editorial extensions

If this is right

  • If EMORe works as claimed, 5D cardiac MRI can tolerate sporadic bulk motion such as coughs, twitches, and deep breaths without requiring navigator echoes, breath holds, or rescanning.
  • Valid but misassigned readouts are corrected rather than thrown away, so the effective acceleration rate is not increased by outlier rejection; only genuinely corrupted readouts land in the outlier bin.
  • The method is a post-acquisition software change: the same k-space data, trajectory, and self-gating signals feed both compressed sensing and EMORe, so existing 5D MRI protocols can adopt it without changing the scan.
  • In the in vivo results, the gains are largest where clinical need is greatest, namely irregular breathing and instructed coughing, so patients with such motion may benefit most.
  • Because the M-step is a generalized EM with only a few ADMM iterations, the framework can be ported to other motion-resolved MRI settings, such as 4D flow, with similar regularization.

Reading between the lines

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

  • A boundary condition the paper does not test: the informative prior pins 85% of each readout's prior probability to the initial self-gating bin, so if a clinical sequence produces systematic binning errors above roughly 15%, EMORe's corrections would be throttled; a natural extension is to make $\alpha_g$ adaptive per readout based on self-gating signal confidence.
  • EMORe corrects discrete bin membership but does not model continuous motion within a bin; combining it with intra-bin non-rigid motion estimation could address residual blur that discrete reassignment cannot fix.
  • The outlier bin could double as a diagnostic signal: the fraction of readouts assigned to it over time is a data-driven record of motion corruption that might flag segments of the acquisition for re-scan or guide prospective gating in future sequences.
  • The Brier-score improvement reported even at 0% simulated outliers suggests soft binning can serve as a post-hoc quality metric for self-gating accuracy, independent of image content.
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

4 major / 3 minor

Summary. The paper proposes EMORe, an expectation-maximization framework for 5D cardiac MRI reconstruction that iteratively refines probabilistic bin assignments and rejects motion-corrupted readouts into a dedicated outlier bin. The E-step computes posterior participation weights from the current image estimate and a bin-assignment prior, and the M-step re-estimates images via sparsity-regularized least squares. The authors validate EMORe against standard compressed sensing in a simulated MRXCAT phantom with controlled bulk motion and in 13 in vivo free-breathing scans, reporting improved PSNR, SSIM, edge sharpness, Brier score, and blinded reader scores.

Significance. If the claims hold, EMORe is a practically valuable drop-in refinement for free-running, free-breathing self-gated 5D cardiac MRI, and the public availability of source code and sample data strengthens reproducibility. The EM derivation is internally consistent, the phantom study provides a quantitative benchmark with ground-truth bin assignments, and the in vivo evaluation includes blinded expert scoring. The main risks are that several load-bearing implementation details are underspecified, in particular the noise-variance estimate used in the E-step, the sensitivity of results to the hand-set binning prior, and a contradiction between the stopping-criterion text and pseudocode. These are fixable with additional analysis and clarification rather than being fundamental flaws.

major comments (4)
  1. [II-A.1, Eq. (1b)-(1c), Table I] The E-step depends critically on the noise standard deviation σ, but the manuscript never states how σ is estimated. Since Table I sets τ = 3σ, the outlier-rejection boundary is determined by σ, and the posterior weights scale as exp(-1/(Lσ²) ||A x - y||²). For in vivo data, especially after coil compression, the noise level is not reported or characterized. The authors should specify a concrete noise-estimation procedure (e.g., from background regions or k-space corners) and, ideally, report sensitivity of the results to σ, because without this the in vivo E-step is not fully reproducible.
  2. [II-A.3, Table I] With αg = 0.85, αo = 0.05, and K = 80, every non-self-gating prior probability equals (1 - 0.85 - 0.05)/79 ≈ 0.00127, so a readout requires a likelihood ratio exceeding roughly 0.85/0.00127 ≈ 670 to be reassigned from its initial self-gating bin. The authors acknowledge in Section II-A.3 that EM is sensitive to initialization, and all hyperparameters were tuned on a single phantom dataset, yet no experiment varies αg or degrades the initial self-gating labels. This leaves the central generalizability claim—that EMORe robustly corrects realistic self-gating inaccuracies—unsecured. A sensitivity analysis over αg and over initial binning accuracy is needed.
  3. [Algorithm 1, line 7; Section II-A.3] The pseudocode stops when the normalized squared image difference is less than η AND t ≥ J, but the text states the stopping criterion is either the maximum number of iterations J or the threshold η, whichever is achieved first. With the AND condition as written, the loop would not terminate at J if the normalized difference has not fallen below η, contradicting the text and potentially affecting runtime and convergence behavior. The condition should be an OR (or the text should be revised to match the pseudocode).
  4. [II-B, Brier score definition] The Brier score is defined as a sum over the K valid bins, while the algorithm assigns posterior mass to K+1 bins including the outlier bin. The manuscript does not specify how corrupted readouts (whose true class is the outlier bin) are represented in the true participation weights w̃, or whether they are excluded from the Brier computation. Without this clarification, the claimed bin-assignment accuracy improvement is ambiguous and may not fully reflect the method's outlier-rejection behavior.
minor comments (3)
  1. [V. Conclusion] There is a typo in the Conclusion: 'implmented' should be 'implemented'.
  2. [III-B, Table II] The text states that blind scoring was performed on 26 cine pairs, while Table II reports per-reviewer means. Please clarify whether the paired t-test was performed across the 26 pairs or across the 13 volunteers, and specify how the per-reviewer values in Table II were aggregated.
  3. [IV. Discussion] The claim of robustness to coughing-induced motion is based on only three volunteers. The authors should state this small-sample limitation explicitly when discussing the in vivo coughing results, rather than presenting the qualitative evidence from Fig. 6 as the primary support.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity: EMORe's improvements are empirical against external ground truth and a compressed-sensing baseline.

full rationale

The paper's central claim is an empirical improvement over compressed sensing on simulated MRXCAT and in vivo data, not an algebraic consequence of its own equations. The EMORe likelihood (Eq. 1), prior (Eq. 4), and M-step objective (Eq. 2) are fixed before the experiments, and the reported PSNR/SSIM/edge sharpness are computed against external ground truth (MRXCAT) or blinded human scoring (in vivo), with CS as an independent comparator. The informative bin prior θ(n,k) is hand-set (αg=0.85), not fitted to the target metric, so the inter-bin correction is not a renamed fit; its strength is a robustness concern, not a circular step. The only self-citation [34] is a preliminary conference version of the same method and is used for attribution, not as evidence. Hyperparameters were tuned on one phantom dataset, which may mildly overstate generalizability, but the evaluation is still external to the derivation chain. No equation in the paper reduces the predicted output to an input by construction. Score 1 reflects a minor preliminary self-citation and tuning concern, not circularity.

Assumptions & free parameters 7 free parameters · 5 assumptions · 0 invented entities

The central claim rests on several domain assumptions about noise statistics, outlier behavior, initial binning quality, and regularization, plus hyperparameters tuned on a single phantom dataset. No new physical entity is introduced.

free parameters (7)
  • αg = 0.85
    Prior probability for the initial SG bin assignment in θ(n,k) (Eq. 4); tuned on one phantom dataset; higher values stabilize but limit bin correction.
  • αo = 0.05
    Prior probability for the outlier bin; tuned on one phantom dataset; controls aggressiveness of outlier rejection.
  • τ = 3σ
    Outlier threshold in the outlier-bin likelihood (Eq. 1c); chosen by hand/tuned; smaller τ rejects more data.
  • σ = not specified
    Noise standard deviation in the likelihood model; known for phantom (30 dB SNR) but estimation method for in vivo data is not described.
  • λs, λc, λr = 2e-2, 10e-2, 6e-2
    Spatial, cardiac, and respiratory TV regularization weights in Eq. 2; tuned on one phantom dataset.
  • I1, I2 = 10, 4
    Inner ADMM iteration counts for initialization and M-step; hand-chosen to balance computation and convergence.
  • J, η = 60, 1e-4
    Outer EM iteration cap and convergence threshold; hand-chosen stopping criteria.
assumptions (5)
  • domain assumption k-space data follow a circularly symmetric Gaussian noise model with known standard deviation σ
    Used in likelihood Eq. (1b) to compute bin-posterior probabilities; if the noise is non-Gaussian or σ misspecified, the soft weights are miscalibrated.
  • domain assumption Bulk-motion-corrupted readouts have a constant likelihood exp(-τ^2/σ^2), independent of the severity of motion
    Eq. (1c) models the outlier bin; this uniform model may not capture varying degrees of corruption.
  • domain assumption The SG-based initial assignments are correct often enough that the informative prior with αg=0.85 guides EM to a reasonable local optimum
    Section II-A.3 acknowledges EM sensitivity to initialization; the prior leans 85% on initial SG binning, so systematic SG errors could trap the solution.
  • domain assumption Anisotropic TV regularization along spatial, cardiac, and respiratory dimensions is an appropriate image prior
    Eq. (2) uses λs, λc, λr; this prior biases the reconstruction toward piecewise-smooth images.
  • domain assumption MRXCAT phantom simulation with PCA/ICA-based self-gating reproduces the error characteristics of real free-running 5D MRI
    The in vivo validation lacks ground truth; the phantom is the only quantitative benchmark for bin assignment accuracy and PSNR/SSIM.

how reviews work

0 comments
Cite this review

Pith. "Pith review of EMORe: Motion-Robust 5D MRI Reconstruction via Expectation-Maximization-Guided Binning Correction and Outlier Rejection." pith.science (2026). https://pith.science/paper/OXMCDHWH

@misc{pith2026250723224,
  author       = {Pith},
  title        = {Pith review of: EMORe: Motion-Robust 5D MRI Reconstruction via Expectation-Maximization-Guided Binning Correction and Outlier Rejection},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/OXMCDHWH}},
  note         = {Machine review of arXiv:2507.23224}
}
read the original abstract

We propose EMORe, an adaptive reconstruction method designed to enhance motion robustness in free-running, free-breathing self-gated 5D cardiac magnetic resonance imaging (MRI). Traditional self-gating-based motion binning for 5D MRI often results in residual motion artifacts due to inaccuracies in cardiac and respiratory signal extraction and sporadic bulk motion, compromising clinical utility. EMORe addresses these issues by integrating adaptive inter-bin correction and explicit outlier rejection within an expectation-maximization (EM) framework, whereby the E-step and M-step are executed alternately until convergence. In the E-step, probabilistic (soft) bin assignments are refined by correcting misassignment of valid data and rejecting motion-corrupted data to a dedicated outlier bin. In the M-step, the image estimate is improved using the refined soft bin assignments. Validation in a simulated 5D MRXCAT phantom demonstrated EMORe's superior performance compared to standard compressed sensing reconstruction, showing significant improvements in peak signal-to-noise ratio, structural similarity index, edge sharpness, and bin assignment accuracy across varying levels of simulated bulk motion. In vivo validation in 13 volunteers further confirmed EMORe's robustness, significantly enhancing blood-myocardium edge sharpness and reducing motion artifacts compared to compressed sensing, particularly in scenarios with controlled coughing-induced motion. Although EMORe incurs a modest increase in computational complexity, its adaptability and robust handling of bulk motion artifacts significantly enhance the clinical applicability and diagnostic confidence of 5D cardiac MRI.

Figures

Figures reproduced from arXiv: 2507.23224 by the authors.

Figure 1
Figure 1. Schematic overview of the proposed EMORe framework. (Left) In the E-step (1), we refine bin participation of readouts to valid motion bins and an outlier bin, given the prior bin participation and current image estimate. (Right) In the M-step (2), we improve the image estimate using the refined bin participation. Both steps are repeated until convergence, resulting in motion-compensated images. Assuming that the pri… view at source ↗
Figure 2
Figure 2. Quantitative comparison between EMORe (red) and CS (blue) reconstructions for 5D MRXCAT phantom study across varying levels of simulated motion outliers (0–70%). Metrics shown include PSNR (dB), SSIM, edge sharpness, and Brier score, averaged across five digital subjects. Error bars represent standard error of the mean. Asterisks indicate statistical significance (p < 0.05) using a paired t-test across five subjects… view at source ↗
Figure 3
Figure 3. Representative short-axis slices at end-expiratory and end-inspiratory states extracted from 5D MRXCAT reconstructions using CS and EMORe under 10%, 20%, and 40% simulated motion outlier levels. Each group shows the static frame, and the corresponding temporal profiles along the x–t and y–t dimensions, extracted at the indicated spatial positions (yellow crosshairs). Arrows highlight motion artifacts and blurring in… view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Respiratory surrogate signals (blue) overlaid with simulated bulk motion intervals (horizontal black bars) and the assignment percentage to the outlier bin for the corresponding readout (vertical red bars), shown for three representative cases with 10%, 20%, and 40% mo…
Figure 5
Figure 5. Figure 5: Representative sagittal cine frames at end-expiratory and end-inspiratory phases for three in vivo volunteers reconstructed using CS and EMORe. Corresponding temporal profiles along the x–t and y–t dimensions are presented for qualitative assessment of temporal consist…
Figure 6
Figure 6. Figure 6: Respiratory surrogate signals (blue) from the three volunteers in [PITH_FULL_IMAGE:figures/full_fig_p009_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

45 extracted references · 43 canonical work pages

  1. [1]

    4D flow cardiovascular magnetic resonance con- sensus statement,

    P. Dyverfeldt et al. , “4D flow cardiovascular magnetic resonance con- sensus statement,” J. Cardiovasc. Magn. Reson. , vol. 17, no. 1, p. 72, 2015

    work page 2015

  2. [2]

    Recent advances in cardiovascular magnetic resonance: techniques and applications,

    M. Salerno, B. Sharif, H. Arheden, A. Kumar, L. Axel, D. Li, and S. Neubauer, “Recent advances in cardiovascular magnetic resonance: techniques and applications,” Circ. Cardiovasc. Imaging , vol. 10, no. 6, p. e003951, 2017

    work page 2017

  3. [3]

    Comparison between three-dimensional navigator-gated whole-heart MRI and two-dimensional cine MRI in quantifying ven- tricular volumes,

    H. W. Goo, “Comparison between three-dimensional navigator-gated whole-heart MRI and two-dimensional cine MRI in quantifying ven- tricular volumes,” Korean J. Radiol. , vol. 19, no. 4, pp. 704–714, 2018

    work page 2018

  4. [4]

    Motion artifacts in MRI: A complex problem with many partial solutions,

    M. Zaitsev, J. Maclaren, and M. Herbst, “Motion artifacts in MRI: A complex problem with many partial solutions,”J. Magn. Reson. Imaging, vol. 42, no. 4, pp. 887–901, 2015

    work page 2015

  5. [5]

    Cardiac-respiratory gating method for magnetic reso- nance imaging of the heart,

    Q. Yuan et al. , “Cardiac-respiratory gating method for magnetic reso- nance imaging of the heart,” Magn. Reson. Med., vol. 43, no. 2, pp. 314– 318, 2000

    work page 2000

  6. [6]

    Respiratory bellows revisited for motion compen- sation: preliminary experience for cardiovascular MR,

    C. Santelli et al. , “Respiratory bellows revisited for motion compen- sation: preliminary experience for cardiovascular MR,” Magn. Reson. Med., vol. 65, no. 4, pp. 1097–1102, 2011

    work page 2011

  7. [7]

    XD-GRASP: golden-angle radial MRI with reconstruction of extra motion-state dimensions using compressed sensing,

    L. Feng, L. Axel, H. Chandarana, K. T. Block, D. K. Sodickson, and R. Otazo, “XD-GRASP: golden-angle radial MRI with reconstruction of extra motion-state dimensions using compressed sensing,” Magn. Reson. Med., vol. 75, no. 2, pp. 775–788, 2016

    work page 2016

  8. [8]

    An auto- mated approach to fully self-gated free-running cardiac and respiratory motion-resolved 5D whole-heart MRI,

    L. Di Sopra, D. Piccini, S. Coppo, M. Stuber, and J. Yerly, “An auto- mated approach to fully self-gated free-running cardiac and respiratory motion-resolved 5D whole-heart MRI,” Magn. Reson. Med. , vol. 82, no. 6, pp. 2118–2132, 2019

    work page 2019

Show all 45 references
  1. [9]

    5D flow MRI: a fully self-gated, free-running frame- work for cardiac and respiratory motion–resolved 3D hemodynamics,

    L. E. Ma et al. , “5D flow MRI: a fully self-gated, free-running frame- work for cardiac and respiratory motion–resolved 3D hemodynamics,” Radiol.: Cardiothorac. Imaging , vol. 2, no. 6, p. e200219, 2020

    2020

  2. [10]

    Fully self-gated free-running 3D cartesian cardiac cine with isotropic whole-heart coverage in less than 2 min,

    T. K ¨ustner et al., “Fully self-gated free-running 3D cartesian cardiac cine with isotropic whole-heart coverage in less than 2 min,” NMR Biomed., vol. 34, no. 1, p. e4409, 2021

    2021

  3. [11]

    Fully self-gated whole-heart 4D flow imaging from a 5-minute scan,

    A. Pruitt et al. , “Fully self-gated whole-heart 4D flow imaging from a 5-minute scan,” Magn. Reson. Med. , vol. 85, no. 3, pp. 1222–1236, 2021

    2021

  4. [12]

    Free-running cardiac and respiratory motion-resolved imaging: A paradigm shift for managing motion in cardiac MRI?,

    R. J. Holtackers and M. Stuber, “Free-running cardiac and respiratory motion-resolved imaging: A paradigm shift for managing motion in cardiac MRI?,” Diagnostics, vol. 14, no. 17, p. 1946, 2024

    1946

  5. [13]

    Self-gated cardiac cine MRI,

    A. C. Larson, R. D. White, G. Laub, E. R. McVeigh, D. Li, and O. P. Simonetti, “Self-gated cardiac cine MRI,” Magn. Reson. Med. , vol. 51, no. 1, pp. 93–102, 2004

    2004

  6. [14]

    5D whole-heart sparse MRI,

    L. Feng et al. , “5D whole-heart sparse MRI,” Magn. Reson. Med. , vol. 79, no. 2, pp. 826–838, 2018

    2018

  7. [15]

    Respiratory and cardiac self-gated free-breathing cardiac cine imaging with multi-echo 3D hybrid radial SSFP acquisition,

    J. Liu, P. Spincemaille, N. C. Codella, T. D. Nguyen, M. R. Prince, and Y . Wang, “Respiratory and cardiac self-gated free-breathing cardiac cine imaging with multi-echo 3D hybrid radial SSFP acquisition,” Magn. Reson. Med. , vol. 63, no. 5, pp. 1230–1237, 2010

    2010

  8. [16]

    Free breathing whole-heart 3D cine MRI with self-gated Cartesian trajectory,

    M. Usman, B. Ruijsink, M. Nazir, G. Cruz, and C. Prieto, “Free breathing whole-heart 3D cine MRI with self-gated Cartesian trajectory,” Magn. Reson. Imaging , vol. 38, pp. 129–137, 2017

    2017

  9. [17]

    Free-breathing whole-heart 3D cine magnetic resonance imaging with prospective respiratory motion compensation,

    M. H. Moghari, A. Barthur, M. E. Amaral, T. Geva, and A. J. Powell, “Free-breathing whole-heart 3D cine magnetic resonance imaging with prospective respiratory motion compensation,” Magn. Reson. Med. , vol. 80, no. 1, pp. 181–189, 2018

    2018

  10. [18]

    Free- breathing whole-heart coronary MRA with 3D radial SSFP and self- navigated image reconstruction,

    C. Stehning, P. B ¨ornert, K. Nehrke, H. Eggers, and M. Stuber, “Free- breathing whole-heart coronary MRA with 3D radial SSFP and self- navigated image reconstruction,” Magn. Reson. Med. , vol. 54, no. 2, pp. 476–480, 2005

    2005

  11. [19]

    3D radial sampling and 3D affine transform-based respiratory motion correc- tion technique for free-breathing whole-heart coronary MRA with 100% imaging efficiency,

    H. Bhat, L. Ge, S. Nielles-Vallespin, S. Zuehlsdorff, and D. Li, “3D radial sampling and 3D affine transform-based respiratory motion correc- tion technique for free-breathing whole-heart coronary MRA with 100% imaging efficiency,” Magn. Reson. Med. , vol. 65, no. 5, pp. 1269–...

    2011

  12. [20]

    Nonrigid autofocus motion correction for coronary MR angiography with a 3D cones trajectory,

    R. R. Ingle et al. , “Nonrigid autofocus motion correction for coronary MR angiography with a 3D cones trajectory,” Magn. Reson. Med. , vol. 72, no. 2, pp. 347–361, 2014

    2014

  13. [21]

    Ac- quisition and visualization of 5D respiratory-resolved cardiac MRI,

    H. H. Wu, D. G. Nishimura, B. S. Hu, and M. V . McConnell, “Ac- quisition and visualization of 5D respiratory-resolved cardiac MRI,” J. Cardiovasc. Magn. Reson. , vol. 14, no. Suppl 1, p. P237, 2012

    2012

  14. [22]

    Ferumoxytol-enhanced free-running 5D whole-heart CMR at 0.55 T,

    X. Sieber et al. , “Ferumoxytol-enhanced free-running 5D whole-heart CMR at 0.55 T,” J. Cardiovasc. Magn. Reson. , vol. 27, 2025

    2025

  15. [23]

    E. K. Weiss et al. , “Respiratory-resolved five-dimensional flow car- diovascular magnetic resonance: In-vivo validation and respiratory- dependent flow changes in healthy volunteers and patients with congeni- tal heart disease,” J. Cardiovasc. Magn. Reson., vol. 26, no. 2, p....

    2024

  16. [24]

    L. Feng, L. Axel, H. Chandarana, K. T. Block, D. K. Sodickson, and R. Otazo, “XD-GRASP: golden-angle radial MRI with reconstruction of 10 THIS WORK HAS BEEN SUBMITTED TO THE IEEE TMI FOR POSSIBLE PUBLICATION. extra motion-state dimensions using compressed sensing,” Magn. Reson...

    2016

  17. [25]

    Motion-robust free-running volumetric cardiovas- cular MRI,

    S. M. Arshad et al. , “Motion-robust free-running volumetric cardiovas- cular MRI,” Magn. Reson. Med. , vol. 92, no. 3, pp. 1248–1262, 2024

    2024

  18. [26]

    Pilot tone–based motion correction for prospective respiratory compensated cardiac cine MRI,

    J. Ludwig, P. Speier, F. Seifert, T. Schaeffter, and C. Kolbitsch, “Pilot tone–based motion correction for prospective respiratory compensated cardiac cine MRI,” Magn. Reson. Med. , vol. 85, no. 5, pp. 2403–2416, 2021

    2021

  19. [27]

    Intra-bin correction and inter-bin compensation of respiratory motion in free-running five-dimensional whole-heart mag- netic resonance imaging,

    C. W. Roy et al. , “Intra-bin correction and inter-bin compensation of respiratory motion in free-running five-dimensional whole-heart mag- netic resonance imaging,” J. Cardiovasc. Magn. Reson. , vol. 26, no. 1, p. 101037, 2024

    2024

  20. [28]

    CINENet: deep learning-based 3D cardiac cine MRI reconstruction with multi-coil complex-valued 4D spatio-temporal convolutions,

    T. K ¨ustner et al. , “CINENet: deep learning-based 3D cardiac cine MRI reconstruction with multi-coil complex-valued 4D spatio-temporal convolutions,” Sci. Rep. , vol. 10, no. 1, p. 13710, 2020

    2020

  21. [29]

    Outlier rejection for diffusion weighted imaging,

    M. Niethammer, S. Bouix, S. Aja-Fern ´andez, C.-F. Westin, and M. E. Shenton, “Outlier rejection for diffusion weighted imaging,” in Interna- tional Conference on Medical Image Computing and Computer-Assisted Intervention, pp. 161–168, Springer, 2007

    2007

  22. [30]

    Structural correlation-based outlier rejection (SCORE) algorithm for arterial spin labeling time series,

    S. Dolui, Z. Wang, R. T. Shinohara, D. A. Wolk, J. A. Detre, and A. D. N. Initiative, “Structural correlation-based outlier rejection (SCORE) algorithm for arterial spin labeling time series,” J. Magn. Reson. Imaging , vol. 45, no. 6, pp. 1786–1797, 2017

    2017

  23. [31]

    A wavelet method for modeling and despiking motion artifacts from resting-state fMRI time series,

    A. X. Patel et al., “A wavelet method for modeling and despiking motion artifacts from resting-state fMRI time series,” Neuroimage, vol. 95, pp. 287–304, 2014

    2014

  24. [32]

    G. J. McLachlan and T. Krishnan, The EM Algorithm and Extensions . John Wiley & Sons, 2008

    2008

  25. [33]

    Fetal cardiac cine imaging using highly accelerated dynamic MRI with retrospective motion correction and outlier rejection,

    J. F. van Amerom et al. , “Fetal cardiac cine imaging using highly accelerated dynamic MRI with retrospective motion correction and outlier rejection,” Magn. Reson. Med., vol. 79, no. 1, pp. 327–338, 2018

    2018

  26. [34]

    EMORe: motion- robust XD-CMR reconstruction using expectation-maximization (EM) algorithm,

    S. M. Arshad, L. C. Potter, X. Lei, and R. Ahmad, “EMORe: motion- robust XD-CMR reconstruction using expectation-maximization (EM) algorithm,” J. Cardiovasc. Magn. Reson. , vol. 27, 2025

    2025

  27. [35]

    Distributed optimization and statistical learning via the alternating direction method of multipliers,

    S. Boyd et al. , “Distributed optimization and statistical learning via the alternating direction method of multipliers,” F ound. Trends Mach. Le. , vol. 3, no. 1, pp. 1–122, 2011

    2011

  28. [36]

    High-dimensional cluster analysis with the masked EM algorithm,

    S. N. Kadir, D. F. Goodman, and K. D. Harris, “High-dimensional cluster analysis with the masked EM algorithm,” Neural Comput. , vol. 26, no. 11, pp. 2379–2394, 2014

    2014

  29. [37]

    Compressed sensing MRI,

    M. Lustig, D. L. Donoho, J. M. Santos, and J. M. Pauly, “Compressed sensing MRI,” IEEE Signal Process. Mag. , vol. 25, no. 2, pp. 72–82, 2008

    2008

  30. [38]

    Image quality assessment: from error visibility to structural similarity,

    Z. Wang, A. C. Bovik, H. R. Sheikh, and E. P. Simoncelli, “Image quality assessment: from error visibility to structural similarity,” IEEE Trans. Image Process. , vol. 13, no. 4, pp. 600–612, 2004

    2004

  31. [39]

    Edge sharpness assessment by parametric modeling: application to magnetic resonance imaging,

    R. Ahmad, Y . Ding, and O. P. Simonetti, “Edge sharpness assessment by parametric modeling: application to magnetic resonance imaging,” Concepts Magn. Reson., Part A , vol. 44, no. 3, pp. 138–149, 2015

    2015

  32. [40]

    Verification of forecasts expressed in terms of probability,

    G. W. Brier, “Verification of forecasts expressed in terms of probability,” Mon. Weather Rev , vol. 78, no. 1, pp. 1–3, 1950

    1950

  33. [41]

    MRXCAT: Realistic numerical phantoms for cardiovascular magnetic resonance,

    L. Wissmann, C. Santelli, W. P. Segars, and S. Kozerke, “MRXCAT: Realistic numerical phantoms for cardiovascular magnetic resonance,” J. Cardiovasc. Magn. Reson. , vol. 16, pp. 1–11, 2014

    2014

  34. [42]

    Technical report (v1. 0)–pseudo-random Cartesian sampling for dynamic MRI,

    M. Joshi, A. Pruitt, C. Chen, Y . Liu, and R. Ahmad, “Technical report (v1. 0)–pseudo-random Cartesian sampling for dynamic MRI,” arXiv preprint arXiv:2206.03630, 2022

    2022 arXiv

  35. [43]

    Cardiac and respiratory motion extraction for MRI using Pilot Tone–a patient study,

    C. Chen et al. , “Cardiac and respiratory motion extraction for MRI using Pilot Tone–a patient study,” Int. J. Card. Imaging , vol. 40, no. 1, pp. 93–105, 2024

    2024

  36. [44]

    Array compression for MRI with large coil arrays,

    M. Buehrer, K. P. Pruessmann, P. Boesiger, and S. Kozerke, “Array compression for MRI with large coil arrays,” Magn. Reson. Med. , vol. 57, no. 6, pp. 1131–1139, 2007

    2007

  37. [45]

    Adaptive recon- struction of phased array MR imagery,

    D. O. Walsh, A. F. Gmitro, and M. W. Marcellin, “Adaptive recon- struction of phased array MR imagery,” Magn. Reson. Med. , vol. 43, pp. 682–690, May 2000

    2000

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.