Pith. sign in

REVIEW 3 major objections 5 minor 52 references

Fast MRI of bones in the knee -- An AI-driven reconstruction approach for adiabatic inversion recovery prepared ultra-short echo time sequences

T0 review · 3 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read The paper claims that an AI-driven iterative reconstruction of undersampled IR-UTE knee MRI can produce bone images that agree closely with a 30-minute reference, with the strongest result at a 5-minute scan time.

desk verdict A solid, honest feasibility study of plug-and-play DnCNN acceleration for 3D IR-UTE knee MRI, but the quantitative tables are missing from the text and the retrospective-to-prospective transfer is plausible rather than proven. read the letter →

arxiv 2506.11771 v1 pith:6CT6P6GQ submitted 2025-06-13 physics.med-ph

Philipp Hans Nunn , Henner Huflage , Jan-Peter Grunz , Philipp Gruschwitz , Oliver Schad , Thorsten Alexander Bley , Johannes Tran-Gia , Tobias Wech This is my paper · ORCID
classification physics.med-ph
keywords kneeMRIadiabaticinversionrecoveryultra-shortechotimeDnCNNLandweberiterationplug-and-playreconstructionundersampledradialdeeplearning
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 authors are trying to establish that MR-based bone imaging of the knee does not require a 30-minute scan. They propose a reconstruction method, S3MOB, in which a denoising convolutional neural network trained on this specific sequence regularizes the image inside each Landweber iteration, and they report that it reconstructs 2.5-, 5-, and 10-minute acquisitions in good agreement with the full reference, with the 5-minute case especially favorable. Quantitative evaluation on four volunteers excluded from training and on one prospectively accelerated scan showed higher structural similarity and peak signal-to-noise ratio than CG-SENSE and than the same iterative loop with a generic denoiser, and sharper edges than applying the network only once. If the claim carries to patients, radiation-free bone assessment of the knee becomes practical at scan times compatible with clinical workflow.

What carries the argument

The load-bearing mechanism is a plug-and-play Landweber loop. Starting from the gridding reconstruction $x_0=A^*y$, each of six iterations applies the MRI forward operator $A$ (non-Cartesian gridding plus coil sensitivities), subtracts the measured data, and feeds the residual back through the adjoint $A^*$ as a data-consistency correction; before the network step, the complex phase is stored and the magnitude slice is normalized, denoised by the DnCNN, weighted by $\sigma=0.15$ against the previous iterate, and rescaled. The custom DnCNN is a residual convolutional network with 482.9k parameters, trained on pairs of CG-SENSE reconstructions of retrospectively undersampled subsets and their full 30-minute references from eight volunteers, so it learns IR-UTE-specific noise suppression. This integration reduces noise without the blur and striation artifacts seen when the same network is applied once to a CG-SENSE image.

What would settle it

Run S3MOB on true accelerated 5-minute IR-UTE knee scans from several volunteers and compare SSIM and PSNR against their matching retrospectively undersampled reconstructions: a systematic, artifact-linked drop in the prospective images would falsify the central claim.

Watch

Extended reading notes

Core claim

The paper's central claim is that S3MOB, six Landweber iterations with a plug-and-play DnCNN as regularizer, reconstructs undersampled 3D IR-UTE knee data in good agreement with the 30-minute reference dataset. On four held-out volunteers, the method achieved higher SSIM and PSNR than CG-SENSE and LW-dIf, with a slight reduction in sharpness relative to those baselines, and higher perceptual sharpness than CG-SENSE-Dn. For volunteer V12, a true 5-minute accelerated scan reconstructed with S3MOB compared favorably with its retrospectively undersampled counterpart and showed no visible streaking. The authors conclude that the method preserves contrast and structural detail while suppressing noise, and that it is poised to make MR-based bone assessment possible in clinically feasible scan times.

Load-bearing premise

The load-bearing premise is that retrospectively thinning a fully sampled 30-minute acquisition reproduces a genuinely accelerated scan, even though the accelerated sequence samples seven spokes around the inversion null while the retrospective subset does not.

Editorial extensions

If this is right

  • A 5-minute knee acquisition reconstructed with S3MOB is the operating point the authors recommend, since it gives high similarity to the 30-minute reference while keeping sharpness above the one-shot denoiser baseline.
  • The same fixed settings, six iterations and a denoising weight of 0.15, performed consistently across 10-, 5-, and 2.5-minute protocols, so the pipeline does not require per-scan parameter tuning.
  • The comparison with LW-dIf shows that a denoiser trained on IR-UTE data is a necessary ingredient: swapping in a generic pretrained denoiser lowers SSIM and PSNR, so the learned denoiser, not the iterative loop alone, carries much of the quality gain.
  • In the one prospectively accelerated scan, the 5-minute image showed no visible streaking, suggesting the method can survive real acceleration rather than only simulated undersampling; the 2.5-minute image stayed usable but showed slight blurring and streaking.

Reading between the lines

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

  • Beyond the paper, the equivalence between prospective and retrospective acceleration could be tested directly by varying the number of spokes acquired per inversion pulse, the parameter the authors identify as the main source of discrepancy.
  • If the 5-minute protocol holds in patients, the same reconstruction pipeline offers a path toward replacing CT for selected knee indications, since IR-UTE-derived bone measures already correlate with bone density.
  • A variational-network version, with the denoiser unrolled through the iterations and the weighting learned per step, is the natural follow-up and would plausibly recover some of the sharpness that the fixed weight of 0.15 sacrifices.
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. The manuscript proposes S3MOB, a Landweber iterative reconstruction with a plug-and-play DnCNN, for accelerating 3D IR-UTE knee MRI. The DnCNN is trained on CG-SENSE reconstructions of retrospectively undersampled data from eight volunteers, paired with CG-SENSE references. The method is evaluated on four held-out volunteers (V9-V12), including one volunteer (V12) with two prospectively accelerated scans, using SSIM, PSNR, NRMSE, and PSI. The authors report improved similarity metrics compared to CG-SENSE and LW-dIf, and improved sharpness compared to CG-SENSE-Dn, especially for a 5-minute acquisition.

Significance. If the reported results hold, the method could make 3D IR-UTE bone imaging clinically feasible within about 5 minutes without ionizing radiation. The use of an open-source Pulseq sequence, the detailed reconstruction pseudo-code, and the held-out evaluation design are strengths. However, the evaluation relies on a single prospective subject and on an untested equivalence between retrospective undersampling and true prospective acceleration; the quantitative tables are also not populated in the submitted text. The claimed improvement therefore remains plausible but not yet fully demonstrated.

major comments (3)
  1. [Section 2.6, Tables 1 and 2] The manuscript refers to Table 1 and Table 2 for the quantitative image quality metrics, but the actual numerical values are absent from the submitted text; only the captions are present. Since the central claims of improved SSIM/PSNR/NRMSE relative to CG-SENSE and LW-dIf rest on these numbers, the evaluation cannot be verified as submitted. The tables must be populated and the corresponding averages and standard deviations reported.
  2. [Section 2.3 and Discussion, third paragraph] The equivalence between retrospectively undersampled and prospectively accelerated acquisitions is a load-bearing assumption. As the authors acknowledge, prospective scans acquire seven spokes per inversion pulse around the null point, producing signal cancellation between spokes before and after the null, whereas retrospective subsets are formed by selecting every Nth projection from the combined trajectory and do not reproduce this intra-TR grouping. The claim that 'mutual averaging can be assumed' over 84k spokes is not quantitatively supported, and no experiment isolates the effect of spoke grouping, noise correlation, or the exact k-space sampling pattern. Because the only prospective evidence comes from a single volunteer (V12), this assumption must be tested or explicitly bounded before the 5-minute clinical feasibility claim can be accepted.
  3. [Section 2.6, Table 2] The prospective evaluation is based on a single volunteer (V12), and the prospectively acquired images are rigidly registered to the reference before computing metrics, while the retrospectively undersampled images are not. Registration improves the similarity metrics but lowers PSI, which the authors attribute to blurring introduced by the registration optimizer. This asymmetric handling makes the prospective-versus-retrospective comparison in Table 2 difficult to interpret, and the single-subject design does not support generalization of the conclusion. At minimum, the unregistered prospective metrics should be reported alongside the registered ones, and the analysis should discuss how registration affects the comparison.
minor comments (5)
  1. [Section 2.4 and Figure 2] The text states that 'the network was trained using only magnitude data', but the Figure 2 caption says 'Input and target slices were normalized and phase-shifted prior to training'; please clarify whether phase information is used in training or not.
  2. [Section 4, first paragraph] The phrase 'S3MOB also exhibits significantly higher PSI values' uses the term 'significantly' without any statistical test; either provide significance testing or rephrase to 'higher' to avoid implying formal inference.
  3. [Section 3, first paragraph] The training duration for R=12 is listed as '45 (R = 12) hours' with the unit 'hours' missing; this should read '45 hours (R = 12)'.
  4. [Section 2.6] The text refers to a '2D rectangular region-of-interest (ROI)' and then to a 'volume-of-interest (VOI)' without clearly defining how the VOI is derived from the 2D ROIs; please specify the relationship.
  5. [Abstract and Section 2.2] The abstract says 'one prospectively accelerated scan' but Section 2.2 describes two prospective acquisitions (7k and 14k projections) for volunteer V12; please reconcile this inconsistency.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: held-out retrospective and prospective evaluations are genuine generalization tests; the acknowledged retrospective/prospective discrepancy is a limitation, not a circular step.

full rationale

The derivation chain is self-contained. The DnCNN was trained on CG-SENSE reconstructions of retrospectively undersampled data paired with CG-SENSE references from volunteers V1-V8, and evaluated on held-out volunteers V9-V12 (Sections 2.3-2.4). Because the test subjects are disjoint from training, the reported SSIM/PSNR/NRMSE values are genuine generalization measurements, not fitted outputs; the hyperparameters (n=6, xi=1, sigma=0.15) are fixed constants and are not selected on the evaluation data. The prospective V12 scan is also from a held-out subject and is compared against a separately acquired 30-minute reference, so the claim is not equivalent to the training target by construction. The Discussion (third paragraph) explicitly acknowledges that prospective scans acquire seven spokes per IR pulse around the null, whereas retrospective subsets select every Nth spoke, and argues the difference is averaged out over 84k spokes; this is a validity limitation of the transfer experiment, but it is disclosed and it does not make the reconstruction reduce to its inputs. The only self-citation (reference [31] in the introductory machine-learning list) is contextual background and is not load-bearing for the S3MOB derivation. No fitted parameter is renamed as a prediction, no uniqueness theorem is imported from the authors' prior work, and no known result is repackaged under new coordinates.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

The central claim rests on standard supervised learning assumptions plus two domain-specific premises: the validity of the CG-SENSE reference and the equivalence of retrospective and prospective undersampling. These are reasonable but not fully validated, especially the latter, which the authors themselves discuss. No invented entities were introduced.

free parameters (3)
  • denoising weighting factor sigma = 0.15
    Chosen by the authors and held constant across all reconstructions; affects the balance between data consistency and denoising. Not fitted to test data, but selection criteria are not fully documented.
  • number of Landweber iterations n = 6
    Chosen by the authors; more iterations increase computational cost and may affect convergence. Held constant across reconstructions.
  • Landweber step size xi = 1
    Fixed to 1 in the pseudo-code; controls the update step in the iterative reconstruction.
assumptions (4)
  • domain assumption The 30-minute CG-SENSE reconstruction serves as a valid reference (ground truth) for training and evaluation.
    Used throughout the paper as the comparison target. If this reference contains artifacts or biases, the network learns them and the metrics are not absolute.
  • ad hoc to paper Retrospective undersampling (selecting every Nth projection) is a valid proxy for true prospective acceleration, including the multi-spoke-per-inversion signal cancellation effect.
    Acknowledged in the Discussion: prospective acquisitions acquire seven projections per IR pulse around TI, leading to signal cancellation that retrospective undersampling may not reproduce. The authors argue mutual averaging over 84k spokes, but this is an assumption.
  • domain assumption The DnCNN trained on magnitude images with phase estimated during reconstruction generalizes to unseen subjects and to prospective acquisitions.
    Required for the central claim to hold beyond the training set. Only four test subjects, one prospective, support this.
  • domain assumption The IR-UTE sequence parameters (TI=64 ms, TR=150 ms) provide adequate long-T2 suppression and bone contrast.
    Taken from prior work [8]; the paper does not independently verify the sequence's contrast behavior in this cohort.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Fast MRI of bones in the knee -- An AI-driven reconstruction approach for adiabatic inversion recovery prepared ultra-short echo time sequences." pith.science (2026). https://pith.science/paper/6CT6P6GQ

@misc{pith2026250611771,
  author       = {Pith},
  title        = {Pith review of: Fast MRI of bones in the knee -- An AI-driven reconstruction approach for adiabatic inversion recovery prepared ultra-short echo time sequences},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/6CT6P6GQ}},
  note         = {Machine review of arXiv:2506.11771}
}
read the original abstract

Purpose: Inversion recovery prepared ultra-short echo time (IR-UTE)-based MRI enables radiation-free visualization of osseous tissue. However, sufficient signal-to-noise ratio (SNR) can only be obtained with long acquisition times. This study proposes a data-driven approach to reconstruct undersampled IR-UTE knee data, thereby accelerating MR-based 3D imaging of bones. Methods: Data were acquired with a 3D radial IR-UTE pulse sequence, implemented using the open-source framework Pulseq. A denoising convolutional neural network (DnCNN) was trained in a supervised fashion using data from eight healthy subjects. Conjugate gradient sensitivity encoding (CG-SENSE) reconstructions of different retrospectively undersampled subsets (corresponding to 2.5-min, 5-min and 10-min acquisition times) were paired with the respective reference dataset reconstruction (30-min acquisition time). The DnCNN was then integrated into a Landweber-based reconstruction algorithm, enabling physics-based iterative reconstruction. Quantitative evaluations of the approach were performed using one prospectively accelerated scan as well as retrospectively undersampled datasets from four additional healthy subjects, by assessing the structural similarity index measure (SSIM), the peak signal-to-noise ratio (PSNR), the normalized root mean squared error (NRMSE), and the perceptual sharpness index (PSI). Results: Both the reconstructions of prospective and retrospective acquisitions showed good agreement with the reference dataset, indicating high image quality, particularly for an acquisition time of 5 min. The proposed method effectively preserves contrast and structural details while suppressing noise, albeit with a slight reduction in sharpness. Conclusion: The proposed method is poised to enable MR-based bone assessment in the knee within clinically feasible scan times.

Figures

Figures reproduced from arXiv: 2506.11771 by the authors.

Figure 1
Figure 1. (A): Overview of 3D IR-UTE sequence. Adiabatic inversion pulse followed by seven projections with the central projection beginning at TI = 64 ms, where magnetization of long T2 is supposed to reach the zero crossing. The interval between two consecutive adiabatic IR pulses is TR = 150 ms. (B): Single excitation pulse and projection. τ = 3.8 ms denotes the time interval between two RF pulses. Ramp sampling and short … view at source ↗
Figure 2
Figure 2. Schematic overview of training data creation and the network. From the acquired 30 [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗
Figure 3
Figure 3. Sagittal knee slice of volunteer V9 to compare CG [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Coronal knee slice of volunteer V9 to compare CG [PITH_FULL_IMAGE:figures/full_fig_p009_4.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

52 extracted references · 39 canonical work pages

  1. [1]

    Chong, K

    L.R. Chong, K. Lee, F.Y. Sim, 3D MRI with CT -like bone contrast – An overview of current approaches and practical clinical implementation, European Journal of Radiology 143 (2021) 109915. https://doi.org/10.1016/j.ejrad.2021.109915

    work page arXiv 2021

  2. [2]

    Y.-J. Ma, S. Jerban, H. Jang, D. Chang, E.Y. Chang, J. Du, Quantitative Ultrashort Echo Time (UTE) Magnetic Resonance Imaging of Bone: An Update, Front. Endocrinol. 11 (2020) 567417. https://doi.org/10.3389/fendo.2020.567417

    work page arXiv 2020

  3. [3]

    Horch, D.F

    R.A. Horch, D.F. Gochberg, J.S. Nyman, M.D. Does, Non -invasive Predictors of Human Cortical Bone Mechanical Properties: T2 -Discriminated 1H NMR Compared with High Resoluti on X- ray, PLoS ONE 6 (2011) e16359. https://doi.org/10.1371/journal.pone.0016359

    work page doi:10.1371/journal.pone.0016359 2011

  4. [4]

    J. Du, M. Carl, M. Bydder, A. Takahashi, C.B. Chung, G.M. Bydder, Qualitative and quantitative ultrashort echo time (UTE) imaging of cortical bone, Journal of Magnetic Re sonance 207 (2010) 304–311. https://doi.org/10.1016/j.jmr.2010.09.013

    work page doi:10.1016/j.jmr.2010.09.013 2010

  5. [5]

    Techawiboonwong, H.K

    A. Techawiboonwong, H.K. Song, F.W. Wehrli, In vivo MRI of submillisecond T2 species with two‐dimensional and three‐dimensional radial sequences and applications to the measurement of cortical bone water, NMR in Biomedicine 21 (2008) 59 –70. https://doi.org/10.1002/nbm.1179

    work page doi:10.1002/nbm.1179 2008

  6. [6]

    Robson, P.D

    M.D. Robson, P.D. Gatehouse, M. Bydder, G.M. Bydder, Magnetic Resonance: An Introduction to Ultrashort TE (UTE) Imaging:, Journal of Computer Assisted Tomogra phy 27 (2003) 825–

    work page 2003

  7. [7]

    J. Du, M. Bydder, A.M. Takahashi, M. Carl, C.B. Chung, G.M. Bydder, Short T2 contrast with three-dimensional ultrashort echo time imaging, Magnetic Resonance Imaging 29 (2011) 470–482. https://doi.org/10.1016/j.mri.2010.11.003

    work page doi:10.1016/j.mri.2010.11.003 2011

  8. [8]

    Y. Ma, Y. Chen, L. Li, Z. Cai, Z. Wei, S. Jerban, H. Jang, E.Y. Chang, J. Du, Trabecular bone imaging using a 3D adiabatic inversion recovery prepared ultrashort TE Cones sequence at 3T, Magn Reson Med 83 (2020) 1640–1651. https://doi.org/10.1002/mrm.28027

    work page doi:10.1002/mrm.28027 2020

Show all 52 references
  1. [9]

    Y. Ma, H. Jang, S. Jerban, E.Y. Chang, C.B. Chung, G.M. Bydder, J. Du, Making the invisible visible—ultrashort echo time magnetic resonance imaging: Technical developments and applications, Applied Physics Reviews 9 (2022) 041303. https://doi.org/10.1063/5.0086459

    2022 doi

  2. [10]

    Chang, J

    E.Y. Chang, J. Du, C.B. Chung, UTE imaging in the musculoskeletal system, Magnetic Resonance Imaging 41 (2015) 870–883. https://doi.org/10.1002/jmri.24713

    2015 doi

  3. [11]

    Rahmer, P

    J. Rahmer, P. Börnert, J. Groen, C. Bos, Three‐dimensional radial ultrashort echo‐time imaging with T 2 adapted sampling, Magnetic Resonance in Med 55 (2006) 1075 –1082. https://doi.org/10.1002/mrm.20868

    2006 doi

  4. [12]

    Rahmer, P

    J. Rahmer, P. Börnert, S.P.M. Dries, Assessment of anterior cruciate ligament reconstruction using 3D ultrashort echo‐time MR imaging, Magnetic Resonance Imaging 29 (2009) 443–448. https://doi.org/10.1002/jmri.21653

    2009 doi

  5. [13]

    Johnson, U

    E.M. Johnson, U. Vyas, P. Ghanouni, K.B. Pauly, J.M. Pauly, Improved cortical bone specificity in UTE MR Imaging, Magnetic Resonance in Med 77 (2017) 684 –695. https://doi.org/10.1002/mrm.26160

    2017 doi

  6. [14]

    H. Lee, X. Zhao, H.K. Song, R. Zhang, S.P. Bartlett, F.W. Wehrli, Rapid dual‐RF, dual‐echo, 3D ultrashort echo time craniofacial imaging: A feasibility study, Magnetic Resonance in Med 81 (2019) 3007–3016. https://doi.org/10.1002/mrm.27625

    2019 doi

  7. [15]

    black -bone

    N. Kamona, B.C. Jones, H. Lee, H.K. Song, C.S. Rajapakse, C.S. Wagner, S.P. Bartlett, F.W. Wehrli, Cranial bone imaging using ultrashort echo -time bone -selective MRI as an alternative to gradient-echo based “black -bone” techniques, Magn Reson Mater Phy 37 (2023) 83 –92. htt...

    2023 doi

  8. [16]

    Kamona, J.J

    N. Kamona, J.J. Ng, Y. Kim, B. -T. D. Vu, A. Vossough, C.S. Wagner, H. Cordray, H. Lee, D.F. Villavisanis, C.S. Rajapakse, S.P. Bartlett, F.W. Wehrli, Craniofacial Imaging of Pediatric Patients by Ultrashort Echo -Time Bone -Selective MRI in Comparison to CT, Academic Radiolog...

    2024 doi

  9. [17]

    Larson, S .M

    P.E.Z. Larson, S .M. Conolly, J.M. Pauly, D.G. Nishimura, Using adiabatic inversion pulses for long‐ T 2 suppression in ultrashort echo time (UTE) imaging, Magnetic Resonance in Med 58 (2007) 952–961. https://doi.org/10.1002/mrm.21341

    2007 doi

  10. [18]

    Afsahi, Y

    A.M. Afsahi, Y. Ma, H. Jang, S. Jerban, C.B. Chung, E.Y. Chang, J. Du, Ultrashort Echo Time Magnetic Resonance Imaging Techniques: Met and Unmet Needs in Musculoskeletal Imaging, Magnetic Resonance Imaging 55 (2022) 1597–1612. https://doi.org/10.1002/jmri.28032

    2022 doi

  11. [19]

    Garwood, L

    M. Garwood, L. DelaBarre, The Return of the Frequency Sweep: Designing Adiabatic Pulses for Contemporary NMR, Journal of Magnetic Resonance 153 (2001) 155 –177. https://doi.org/10.1006/jmre.2001.2340

    2001

  12. [20]

    Nazaran, M

    A. Nazaran, M. Carl, Y. Ma, S. Jerban, Y. Zhu, X. Lu, J. Du, E.Y. Chang, Three-dimensional adiabatic inversion recovery prepared ultrashort echo time cones (3D IR -UTE-Cones) imaging of cortical bone in the hip, Magnetic Resonance Imaging 44 (2017) 60 –64. https://doi.org/10.1...

    2017 doi

  13. [21]

    Y. Ma, J. West, A. Nazar an, X. Cheng, H. Hoenecke, J. Du, E.Y. Chang, Feasibility of using an inversion-recovery ultrashort echo time (UTE) sequence for quantification of glenoid bone loss, Skeletal Radiol 47 (2018) 973–980. https://doi.org/10.1007/s00256-018-2898-4

    2018 doi

  14. [22]

    Ma nhard, R.A

    M.K. Ma nhard, R.A. Horch, D.F. Gochberg, J.S. Nyman, M.D. Does, In Vivo Quantitative MR Imaging of Bound and Pore Water in Cortical Bone, Radiology 277 (2015) 221 –229. https://doi.org/10.1148/radiol.2015140336

    2015 doi

  15. [23]

    Jerban, Y

    S. Jerban, Y. Ma, S. Alenezi, D. Moazamian, J. Athertya, H. Jang, E. Dorthe, D. Dlima, G. Woods, C.B. Chung, E.Y. Chang, J. Du, Ultrashort Echo Time (UTE) MRI porosity index (PI) and suppression ratio (SR) correlate with the cortical bone microstructural and mechanical propert...

    2023

  16. [24]

    Liu, J.-D

    J. Liu, J.-D. Chen, P. Li, J.-W. Liao, J.-X. Feng, Z.-Y. Chen, Z.-Y. Cai, W. Li, X.-J. Chen, Z.-H. Su, H. Lu, S.-L. Li, Y. -J. Ma, Comprehensive assessment of osteoporosis in lumbar spine us ing compositional MR imaging of trabecular bone, Eur Radiol 33 (2022) 3995 –4006. http...

    2022 doi

  17. [25]

    Lustig, D

    M. Lustig, D. Donoho, J.M. Pauly, Sparse MRI: The application of compressed sensing for rapid MR imaging, Magnetic Resonance i n Med 58 (2007) 1182 –1195. https://doi.org/10.1002/mrm.21391

    2007 doi

  18. [26]

    Athertya, Y

    J.S. Athertya, Y. Ma, A. Masoud Afsahi, A.F. Lombardi, D. Moazamian, S. Jerban, S. Sedaghat, H. Jang, Accelerated Quantitative 3D UTE-Cones Imaging Using Compressed Sensing, Sensors 22 (2022) 7459. https://doi.org/10.3390/s22197459

    2022 doi

  19. [27]

    Sriram, J

    A. Sriram, J. Zbontar, T. Murrell, A. Defazio, C.L. Zitnick, N. Yakubova, F. Knoll, P. Johnson, End- to-End Variational Networks for Accelerated MRI Reconstruction, (2020). http://arxiv.org/abs/2004.06688 (accessed August 5, 2024)

    2020 arXiv

  20. [28]

    Muckley, B

    M.J. Muckley, B. Riemenschneider, A. Radmanesh, S. Kim, G. Jeong, J. Ko, Y. Jun, H. Shin, D. Hwang, M. Mostapha, S. Arberet, D. Nickel, Z. Ramzi, P. Ciuciu, J. -L. Starck, J. Teuwen, D. Karkalousos, C. Zhang, A. Sriram, Z. Huang, N. Yakubova, Y.W. Lui, F. Knoll, Results of the...

    2021

  21. [29]

    Aggarwal, M.P

    H.K. Aggarwal, M.P. Mani, M. Jacob, MoDL: Mo del-Based Deep Learning Architecture for Inverse Problems, IEEE Trans. Med. Imaging 38 (2019) 394 –405. https://doi.org/10.1109/TMI.2018.2865356

    2019

  22. [30]

    Ramzi, J

    Z. Ramzi, J. -L. Starck, P. Ciuciu, Density Compensated Unrolled Networks for Non -Cartesian MRI Reconstruction, (2021). http://arxiv.org/abs/2101.01570 (accessed August 5, 2024)

    2021 arXiv

  23. [31]

    Kleineisel, J.F

    J. Kleineisel, J.F. Heidenreich, P. Eirich, N. Petri, H. Köstler, B. Petritsch, T.A. Bley, T. Wech, Real- time cardiac MRI using an undersampled spiral k-space trajectory and a reconstruction based on a variational network, Magnetic Resonance in Medicine 88 (2022) 2167 –2178. ...

    2022 doi

  24. [32]

    Baker, V

    R.R. Baker, V. Muthurangu, M. Rega, S.B. Walsh, J.A. Steeden, Rapid 2D 23Na MRI of the calf using a denoising convolutio nal neural network, Magnetic Resonance Imaging 110 (2024) 184–194. https://doi.org/10.1016/j.mri.2024.04.027. 16

    2024 doi

  25. [33]

    Kaye, E.A

    E.A. Kaye, E.A. Aherne, C. Duzgol, I. Häggström, E. Kobler, Y. Mazaheri, M.M. Fung, Z. Zhang, R. Otazo, H.A. Vargas, O. Akin, Accelerating Pr ostate Diffusion -weighted MRI Using a Guided Denoising Convolutional Neural Network: Retrospective Feasibility Study, Radiology: Artif...

    2020 doi

  26. [34]

    Layton, S

    K.J. Layton, S. Kroboth, F. Jia, S. Lit tin, H. Yu, J. Leupold, J. -F. Nielsen, T. Stöcker, M. Zaitsev, Pulseq: A rapid and hardware -independent pulse sequence prototyping framework: Rapid Hardware-Independent Pulse Sequence Prototyping, Magn. Reson. Med. 77 (2017) 1544 –

    2017

  27. [35]

    Carl, G.M

    M. Carl, G.M. Bydder, J. Du, UTE imaging with simultaneous water and fat signal suppression using a time -efficient multispoke inversion recovery pulse sequence: Optimized Method for Long T2 Signal Suppression, Magn. Reson. Med. 76 (2016 ) 577 –582. https://doi.org/10.1002/mrm.25823

    2016 doi

  28. [36]

    González, Measurement of Areas on a Sphere Using Fibonacci and Latitude –Longitude Lattices, Math Geosci 42 (2010) 49–64

    Á. González, Measurement of Areas on a Sphere Using Fibonacci and Latitude –Longitude Lattices, Math Geosci 42 (2010) 49–64. https://doi.org/10.1007/s11004-009-9257-x

    2010 doi

  29. [37]

    Zur, M.L

    Y. Zur, M.L. Wood, L.J. Neuringer , Spoiling of transverse magnetization in steady‐state sequences, Magnetic Resonance in Med 21 (1991) 251 –263. https://doi.org/10.1002/mrm.1910210210

    1991 doi

  30. [38]

    Muckley, R

    M.J. Muckley, R. Stern, T. Murrell, F. Knoll, TorchKbNufft: A High -Level, Hardware-Agnostic Non-Uniform Fast Fourier Transform, (2020). https://github.com/marcsous/nufft_3d

    2020

  31. [39]

    https://github.com/marcsous/nufft_3d (accessed November 7, 2024)

    MATLAB implementation of Non -Uniform Fast Fourier Transform in 3D, (n.d.). https://github.com/marcsous/nufft_3d (accessed November 7, 2024)

    2024

  32. [40]

    Pruessmann, M

    K.P. Pruessmann, M. Weiger, P. Börnert, P. Boesiger, Advances in sensitivity encoding with arbitrary k ‐space trajectories, Magnetic Resonance in Med 46 (2001) 638 –651. https://doi.org/10.1002/mrm.1241

    2001 doi

  33. [41]

    Walsh, A.F

    D.O. Walsh, A.F. Gmitro, M.W. Marcellin, Adaptive reconstruction of phased array MR imagery, Magn. Reson. Med. 43 (2000) 682 –690. https://doi.org/10.1002/(SICI)1522 - 2594(200005)43:5<682::AID-MRM10>3.0.CO;2-G

    2000 doi

  34. [42]

    Zhang, W

    K. Zhang, W. Zuo, Y. Chen, D. Meng, L. Zhang, Beyond a Gaussian Denoiser: Residual Learning of Deep CNN for Image Den oising, IEEE Trans. on Image Process. 26 (2017) 3142 –3155. https://doi.org/10.1109/TIP.2017.2662206

    2017

  35. [43]

    K. He, X. Zhang, S. Ren, J. Sun, Deep Residual Learning for Image Recognition, in: 2016 IEEE Conference on Computer Vision and Pattern Recognition (CV PR), IEEE, Las Vegas, NV, USA, 2016: pp. 770–778. https://doi.org/10.1109/CVPR.2016.90

    2016 doi

  36. [44]

    Singh, A

    G. Singh, A. Mittal, N. Aggarwal, ResDNN: deep residual learning for natural image denoising, IET Image Process 14 (2020) 2425–2434. https://doi.org/10.1049/iet-ipr.2019.0623

    2020

  37. [45]

    Kingma, J

    D.P. Kingma, J. Ba, Adam: A Method for Stochastic Optimization, (2017). http://arxiv.org/abs/1412.6980 (accessed October 30, 2024)

    2017 arXiv

  38. [46]

    Landweber, An Iteration Formula for Fredholm Integral Equations of the First Kind, American Journal of Mathematics 73 (1951) 615

    L. Landweber, An Iteration Formula for Fredholm Integral Equations of the First Kind, American Journal of Mathematics 73 (1951) 615. https://doi.org/10.2307/2372313

    1951 doi

  39. [47]

    Kamilov, C.A

    U.S. Kamilov, C.A. Bouman, G.T. Buzzard, B. Wohlberg, Plug-and-Play Methods for Integrating Physical and Learned Models in Computational Imaging: Theory, algorithms, and applications, IEEE Signal Process. Mag. 40 (2023) 85–97. https://doi.org/10.1109/MSP.2022.3199595

    2023

  40. [48]

    Feichtenhofer, H

    C. Feichtenhofer, H. Fassold, P. Schallauer, A Perceptual Image Sharpness Metric Based on Local Edge Gradient Analysis, IEEE Signal Process. Lett. 20 (2013) 379 –382. https://doi.org/10.1109/LSP.2013.2248711

    2013

  41. [49]

    L. Wan, W. Zhao, Y. Ma, S. Jerban, A.C. Searleman, M. Carl, E.Y. Chang, G. Tang, J. Du, Fast quantitative 3D ultrashort echo time MRI of cortical bone using extended cones sampling, Magnetic Resonance in Med 82 (2019) 225–236. https://doi.org/10.1002/mrm.27715

    2019 doi

  42. [50]

    Hammernik, T

    K. Hammernik, T. Klatzer, E. Kobler, M.P. Recht, D.K. Sodickson, T. Pock, F. Knoll, Learning a variational network for reconstruction of accelerated MRI data, Magnetic Resonance in Med 79 (2018) 3055–3071. https://doi.org/10.1002/mrm.26977. 17 Supplementary Material Supplement...

    2018 doi

  43. [846]

    https://doi.org/10.1097/00004728-200311000-00001

    doi

  44. [1552]

    https://doi.org/10.1002/mrm.26235

    doi

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.