Pith. sign in

REVIEW 4 major objections 5 minor 35 references

Adaptive Plane Reformatting for 4D Flow MRI using Deep Reinforcement Learning

T0 review · 4 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read This paper argues that a reinforcement-learning agent can reformat 4D flow MRI planes independently of scan orientation by acting in a plane-local coordinate system, and that the resulting flow measurements match expert observers as…

desk verdict AdaPR is a solid incremental advance in DRL plane reformatting; fix the Table 5 inconsistency and tone down the 'arbitrary orientations' claim. read the letter →

arxiv 2506.00727 v2 pith:G2Y5ARLY submitted 2025-05-31 cs.LG cs.CV

Javier Bisbal , Julio Sotelo , Maria I Valdés , Pablo Irarrazaval , Marcelo E Andia , Julio García , José Rodriguez-Palomarez , Francesca Raimondi
show 2 more authors
Cristián Tejos Sergio Uribe
This is my paper
classification cs.LGcs.CV
keywords deepreinforcementlearningplanereformatting4DflowMRIlocalcoordinatesystemA3Cquantificationcardiovascularorientationinvariance
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

AdaPR tries to make automated plane reformatting for 4D flow MRI independent of how a scan was acquired: instead of moving the plane in the scanner's fixed coordinate axes, a deep reinforcement-learning agent rotates and translates the plane along the plane's own local axes ($\vec n_t$, $\vec w^1_t$, $\vec w^2_t$). On 88 multi-vendor scans, the paper reports a mean angular error of $6.32^\circ \pm 4.15^\circ$ and a distance error of $3.40 \pm 2.75$ mm relative to the manual planes of one expert, outperforming a global-coordinate DQN baseline and remaining near-constant when volumes are rotated and translated by 5–25 degrees/millimetres. Flow measured on AdaPR planes showed no significant difference from two manual observers, with $R^2 = 0.972$ and $0.968$, comparable to inter-observer agreement ($R^2 = 0.969$). If correct, this makes automated plane placement usable across scanners and institutions without volume pre-registration, by removing the orientation assumption built into earlier DRL view-planning agents.

What carries the argument

The load-bearing mechanism is the local orthonormal basis $\{\vec n_t, \vec w^1_t, \vec w^2_t\}$: the first axis is the current plane normal and the other two span the plane, with a state sampled as a 3D sub-volume along these axes plus auxiliary lines along $\vec w^1_t$ and $\vec w^2_t$ that preserve orientation information. Actions in Eq. (4) are rotations of the normal around the two in-plane axes and translations of the center along the three local axes, which makes the policy equivariant to rigid transformations of the whole volume. The agent's reward is the temporal decrease of the cost $C(t)$ in Eq. (7), with a terminal bonus of 3 when the angle falls below $3^\circ$ and the distance below 2 mm, and training uses the on-policy A3C algorithm with an LSTM between consecutive states. This combination is what lets the plane be steered in arbitrary scanner coordinates without pre-registration.

What would settle it

Run AdaPR on volumes with known ground-truth planes, obtained from a flow phantom with a calibrated double-oblique cut or from synthetic 4D flow data, and compare the recovered planes with the known targets over many orientations. If the angular error systematically exceeds the reported $6.32^\circ$ or depends on the volume's orientation, the orientation-independence claim would be contradicted; the same test with a consensus of several experts as reference would check the manual ground truth assumption.

Watch

Extended reading notes

Core claim

The paper's central claim is that the spatial alignment assumption in prior DRL plane-reformatting methods can be removed by expressing both the state and the actions in a local coordinate system attached to the current plane. At each step, the agent's state is a sub-volume sampled along the local basis $\{\vec n_t, \vec w^1_t, \vec w^2_t\}$ centered at $\vec P_t$, and its actions are a rotation of the normal around $\vec w^1_t$, a rotation around $\vec w^2_t$, and translations along $\vec w^1_t$, $\vec w^2_t$, and $\vec n_t$. Because the same action produces the same relative change regardless of the volume's global orientation, the learned policy transfers to arbitrarily posed volumes. The paper demonstrates this with an A3C implementation (AdaPR) trained against a cost $C(t) = (1 - \vec n_t \cdot \vec n_T / (\|\vec n_t\| \|\vec n_T\|)) + \lambda d(P_t,P_T)$ that balances angular and distance errors, and reports that AdaPR reaches $6.32^\circ \pm 4.15^\circ$ and $3.40 \pm 2.75$ mm against expert planes, that its error stays within $1.5^\circ$ and 1.1 mm of baseline under rigid perturbations, and that flow quantification matches inter-observer performance.

Load-bearing premise

Observer 1's manually placed planes are used as the ground truth in the reward function and as the evaluation reference, so the reported plane accuracy and the 'comparable to observers' flow claim assume that those manual planes are correct and representative of clinical needs.

Editorial extensions

If this is right

  • AdaPR removes the need to pre-register test volumes to a common spatial framework, so a single trained model can reformat planes in scans acquired in arbitrary orientations and positions from different vendors.
  • Flow computed from AdaPR planes is statistically indistinguishable from expert flow measurements, with $R^2 = 0.972$ against observer 1 and $R^2 = 0.968$ against observer 2 versus inter-observer $R^2 = 0.969$.
  • Under rigid rotations and translations of 5–25 degrees and millimetres, AdaPR's average error stays within $1.5^\circ$ and 1.1 mm of its unperturbed performance, whereas the global-coordinate DQN baseline degrades sharply.
  • The on-policy A3C formulation converges where the off-policy DQN did not in the larger local-coordinate state space, suggesting that on-policy methods are the practical choice for this task.
  • Because accuracy held across healthy volunteers and congenital heart disease patients from multiple scanners, the method is expected to generalize to unseen acquisitions of similar anatomy.

Reading between the lines

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

  • A stricter falsification of the invariance claim would be a synthetic ground-truth test, where the target plane is generated by a known double-oblique transform of a reference plane; the paper's evaluation only compares against manual planes.
  • The local-coordinate policy is a general navigation strategy that should transfer to other imaging modalities and other volumetric targeting tasks, such as valve tracking or biopsy planning, but this is an extrapolation beyond the paper's evidence.
  • The reported flow agreement may be forgiving of angular errors because flow is insensitive to orientation below roughly $25^\circ$, so the more meaningful clinical test is whether AdaPR's angular performance holds on a larger patient cohort with complex anatomy.
  • An explicit 'orientation stress test' reporting per-voxel anatomy alignment after rotating the same volume by arbitrary angles would quantify what the paper's axis-aligned perturbations only approximate.
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 / 5 minor

Summary. This paper presents AdaPR, a deep reinforcement learning framework for automated plane reformatting in 4D flow MRI. The key idea is to replace the global coordinate system used by prior DRL view-planning methods with a local, plane-centered coordinate system, so that the agent's rotations and translations are defined relative to the current plane rather than the scanner axes. The method is trained with the A3C algorithm using a reward that penalizes angular and distance errors to a target plane placed by an expert observer. Experiments on 88 multi-vendor, multi-institution datasets with four-fold cross-validation report mean angular errors of 6.32° +/- 4.15° and distance errors of 3.40 +/- 2.75 mm relative to Observer 1, with better performance than two VanillaPR (global-coordinate) DRL baselines and maintained accuracy under simulated rigid transformations. Flow measurements from AdaPR planes correlate strongly with both manual observers (R^2 = 0.972 and 0.968), close to inter-observer agreement (R^2 = 0.969).

Significance. If the reported results hold, AdaPR addresses a genuine practical limitation of current DRL-based plane reformatting: the requirement that test volumes share the spatial alignment of the training data. The local-coordinate formulation is simple, conceptually clear, and directly motivated by the multi-scanner, multi-orientation setting of 4D flow MRI. The empirical study is substantial for the field, including patients with congenital heart disease, multiple scanner vendors, sensitivity analyses, and flow-based validation against two observers. The open-source implementation is a further strength. The significance is tempered, however, by the fact that the evaluation standard is a single observer's manual planes used both for training and for the headline accuracy numbers; the paper also contains a clear tabular inconsistency and uses a statistically questionable cross-study comparison. These issues require revision before the claims can be fully accepted.

major comments (4)
  1. [§2.3.2, Eq. (7), §3.1.4, §4.3] The target planes placed by Observer 1 serve both as the training reward in Eq. (7) and as the reference for the reported errors (6.32°±4.15°, 3.40±2.75 mm) and the flow correlation analysis. The reported 'accuracy' is therefore a measure of agreement with a single, unvalidated annotation standard. The inter-observer comparison (O1 vs O2) mitigates random label noise, but both observers used the same PC-MRA/ParaView protocol, so a systematic bias common to both would be inherited by AdaPR and would inflate both the geometric and flow agreement claims. Section 4.3 acknowledges the reliance on manual labels, but no independent validation (phantom with known plane geometry, multi-observer consensus, or synthetic planes) is provided. This limitation directly affects the absolute-accuracy component of the central claim; please add such a comparison or explicitly reframe the results as agreement with Observer 1's protocol.
  2. [Table 5 and §3.1.4] Table 5 has a serious internal inconsistency: for every row, the 'All' column exactly repeats the LPA column (e.g., volunteers angular 5.96±4.16, patients 7.47±3.95, inter-observer volunteers 4.49±4.25). This is not a valid aggregate of the AAo, PA, RPA, and LPA columns. In addition, the text in §3.1.4 reports patient errors of 7.31°±3.98° and 4.36±3.58 mm, which do not match the patient 'All' entries 7.47°±3.95° and 4.41±3.59 mm in Table 5. These discrepancies need to be corrected and the subgroup analysis rechecked, because they directly affect the reported patient-versus-volunteer comparison.
  3. [§2.3.3 and §3.1.3] The VanillaPR baselines are described as 'trained on data that was pre-aligned to the same orientation and position via rigid registration', but the paper does not state whether the same rigid registration is applied to the test volumes before evaluation, and if so, how it interacts with the sensitivity analysis in §3.1.3, which applies rotations and translations to the input volumes. If the VanillaPR test pipeline omits the pre-registration step, the degradation shown in Figure 2 may reflect a mismatch between training and test conditions rather than an intrinsic limitation of global coordinates; if it includes pre-registration, the perturbation experiment is not a fair test of the full pipeline. Please specify the exact inference-time preprocessing for each algorithm and, ideally, evaluate VanillaPR with its full intended pipeline (including pre-registration) under the same transformations.
  4. [§4.1 and Table 7] The comparison with non-DRL methods in Table 7 is based on two-sample t-tests applied to published means, standard deviations, and sample sizes from different studies. This is statistically inappropriate because the data come from different populations, acquisition protocols, vessel subsets, and evaluation pipelines, and the test ignores within-study correlation and confounding. The resulting statement that AdaPR significantly 'outperforms' the atlas-based and 3D-CNN methods is therefore not supported by the presented analysis. Please either provide a rigorous matched comparison on the same data or rephrase these results as descriptive benchmark comparisons.
minor comments (5)
  1. [§2.1.2] Please clarify how the auxiliary lines are encoded into the state (e.g., as additional input channels or as intensity overlays) and how the network uses them; the current description leaves the implementation ambiguous.
  2. [§2.4.1] The ART ANOVA is applied to repeated measurements from the same subjects across vessels and comparison types; the paper should state whether the model accounts for within-subject correlation or justify why the independence assumption is acceptable.
  3. [§4.3] The phrase '(Table 5 Appendix)' appears to be a typo; it should refer to Table 5 or a proper appendix table. Also, the text in §4.3 quotes patient errors as 7.47° and 4.41 mm, which conflicts with the 7.31° and 4.36 mm values in §3.1.4; please reconcile.
  4. [§2.1.4] Equation (8) defines the reward as a finite difference of the cost, and a terminal reward of 3 is added separately; it would be helpful to state explicitly that the total reward is r_t + 3 at terminal states and how this is incorporated in the advantage calculation, Eq. (11).
  5. [§2.1.3] The action set in Eq. (4) includes rotations about w1 and w2 but no action that rotates the plane about its normal n. If in-plane orientation is irrelevant to the flow measurements, please state that explicitly; otherwise clarify how the auxiliary lines in §2.1.2 are used in the reward or network input.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: AdaPR is trained against manual observer planes, and the primary error metric is the same geometric cost, but the reported numbers come from held-out cross-validation folds and are corroborated by an observer not used in training and by flow measurements that are not part of the reward.

full rationale

The paper's claimed derivation is an empirical RL training pipeline, not a formal derivation that reduces to its inputs. The reward in Eq. (7) (C(t) = (1 - n_t·n_T) + λ d(P_t,P_T)) uses Observer 1's planes as the target, and the evaluation in Sections 2.4.1/3.1 uses the same angular and distance errors relative to that observer. This shared reference is a supervised-learning convention rather than circularity: the four models are trained on 52 of 88 scans each and tested on held-out folds (Section 2.3.1), so the reported 6.32° ± 4.15° and 3.40 ± 2.75 mm are not fit residuals. Model selection uses validation cost, not test outcomes. The paper also provides external anchors: Observer 2 was not used for training (`We used the planes placed by the first observer to train the DRL algorithms and the planes placed by the second observer to assess inter-observer variability`), AdaPR-vs-O2 agreement is similar (R²=0.968), and the flow comparisons (Section 2.5, Tables 5-6) are computed from Chan-Vese segmentations and actual flow values, which are not part of the reward function; the paper even finds significantly larger AdaPR angular errors than inter-observer errors for most vessels, showing the metric is not forced. The limitations section (4.3) explicitly acknowledges dependence on manual label quality, which is a validity caveat, not a circular step. Citations to the authors' prior work are background or methodological context only and are not load-bearing for the main result; no uniqueness theorem or ansatz is imported from self-citations. The central accuracy and flow claims therefore have independent content beyond their training inputs.

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

The central claim rests on standard ML assumptions: an unbiased reference standard, policy generalization, and heuristic preprocessing. The only hand-chosen numeric parameters are training hyperparameters (lambda, step limits, subvolume size); none are fitted to test data. No new physical entities are introduced.

free parameters (4)
  • lambda (cost function weight) = 0.025
    Balances angular vs distance error in Eq. 7; chosen empirically in Section 2.3.4.
  • omega_max / d_max (max rotation/translation per action) = 5 degrees / 5 mm
    Maximum step sizes for plane updates; empirically chosen for smooth transitions (Table 2).
  • t_max (max steps per episode) = 100
    Episode length limit; chosen to ensure convergence (Table 2).
  • State sub-volume dimensions (D,H,W) = 62 x 168 x 168 mm
    Subvolume size selected to cover most structures in the reformatted view (Section 2.3.4).
assumptions (4)
  • domain assumption Manual planes placed by observers are a valid reference standard for plane reformatting and flow quantification.
    Target planes (n_T, P_T) used in the reward (Eq. 7) and in evaluation come from observer 1 and 2 annotations (Section 2.3.2). If these annotations are biased, the reported accuracy and flow agreement are biased accordingly.
  • domain assumption A3C with the described CNN/LSTM architecture learns a policy that generalizes across subjects and vendors.
    The method relies on standard deep RL generalization from 52 training scans to unseen data; no convergence guarantee is given.
  • domain assumption Preprocessing preserves vascular information needed for navigation.
    The divergence-weighted PC-MRA (Eq. 2) and CLAHE are heuristic; the agent's state depends on them retaining relevant anatomy.
  • standard math Trilinear interpolation and rigid rotations are valid numerical operations.
    Standard numerical methods used for resampling and transformation; assumed exact enough for the application.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Adaptive Plane Reformatting for 4D Flow MRI using Deep Reinforcement Learning." pith.science (2026). https://pith.science/paper/G2Y5ARLY

@misc{pith2026250600727,
  author       = {Pith},
  title        = {Pith review of: Adaptive Plane Reformatting for 4D Flow MRI using Deep Reinforcement Learning},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/G2Y5ARLY}},
  note         = {Machine review of arXiv:2506.00727}
}
read the original abstract

Background and Objective: Plane reformatting for four-dimensional phase contrast MRI (4D flow MRI) is time-consuming and prone to inter-observer variability, which limits fast cardiovascular flow assessment. Deep reinforcement learning (DRL) trains agents to iteratively adjust plane position and orientation, enabling accurate plane reformatting without the need for detailed landmarks, making it suitable for images with limited contrast and resolution such as 4D flow MRI. However, current DRL methods assume that test volumes share the same spatial alignment as the training data, limiting generalization across scanners and institutions. To address this limitation, we introduce AdaPR (Adaptive Plane Reformatting), a DRL framework that uses a local coordinate system to navigate volumes with arbitrary positions and orientations. Methods: We implemented AdaPR using the Asynchronous Advantage Actor-Critic (A3C) algorithm and validated it on 88 4D flow MRI datasets acquired from multiple vendors, including patients with congenital heart disease. Results: AdaPR achieved a mean angular error of 6.32 +/- 4.15 degrees and a distance error of 3.40 +/- 2.75 mm, outperforming global-coordinate DRL methods and alternative non-DRL methods. AdaPR maintained consistent accuracy under different volume orientations and positions. Flow measurements from AdaPR planes showed no significant differences compared to two manual observers, with excellent correlation (R^2 = 0.972 and R^2 = 0.968), comparable to inter-observer agreement (R^2 = 0.969). Conclusion: AdaPR provides robust, orientation-independent plane reformatting for 4D flow MRI, achieving flow quantification comparable to expert observers. Its adaptability across datasets and scanners makes it a promising candidate for medical imaging applications beyond 4D flow MRI.

Figures

Figures reproduced from arXiv: 2506.00727 by the authors.

Figure 1
Figure 1. Overview of the adaptive plane-reformatting framework (AdaPR) and reformatted planes. A. Asynchronous Advantage Actor–Critic (A3C) architecture: An initial stack of volumes goes through three 3D convolutional layers and 1 fully-connected layer. The output of this layer feeds an LSTM cell with the previous LSTM hidden and cell state. Next, the Actor layers compute 𝜇𝑎 and 𝜎𝑎 to estimate the policy 𝜋(𝑠, 𝑎) and the Crit… view at source ↗
Figure 2
Figure 2. Sensitivity of plane-reformatting accuracy to rigid transformations of the input volumes. Heatmaps summarize average performance across all test volumes and planes when the same scans are perturbed by rotations of 5°, 15°, 25° and translations of 5, 15, 25 mm along all axes. Top panels report angular error (°); bottom panels report distance error (mm), both computed between the automated planes and the reference man… view at source ↗
Figure 3
Figure 3. Distribution of plane reformatting metrics for AdaPR. Whiskers range from the 2.5th to the 97.5th percentile. The line represents the median, and the cross represents the mean. (A) Angular error (◦ ). (B) Distance error (mm). **, ***, and **** represent ART-C post-hoc contrasts with significance levels < 10−2 ,< 10−3, and < 10−4, respectively. : Preprint submitted to Elsevier Page 18 of 16 [PITH_FULL_IMAGE:figures/… view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Examples of plane reformatting. Planes placed by observer 1 (O1), observer 2 (O2), and AdaPR planes are shown in yellow, green, and red, respectively. The left-hand column depicts a healthy volunteer; the right-hand column shows a patient with a bicuspid aortic valve (…
Figure 5
Figure 5. Figure 5: Flow measurements (L/min) at four vessels (AAo: ascending aorta; PA: pulmonary artery; RPA: right pulmonary artery; LPA: left pulmonary artery) for AdaPR, Observer 1, and Observer 2. Box plots show median (line), mean (+), interquartile range (box),and whiskers range f…
Figure 6
Figure 6. Figure 6: Bland–Altman plots showing percentage flow differences for each vessel (AAo, PA, RPA, LPA). The first row compares AdaPR vs Observer 1, the second row AdaPR vs Observer 2, and the third row Observer 1 vs Observer 2. The solid blue line indicates the mean bias, and the …

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

35 extracted references · 34 canonical work pages

  1. [1]

    Sotelo, I

    J. Sotelo, I. Valverde, D. Martins, D. Bonnet, N. Boddaert, K. Pushparajan, S. Uribe, F. Raimondi, Impact of aortic arch curvature in flow haemodynamics in patients with transposition of the great arteries after arterial switch operation, European Heart Journal - Cardiovascular Imaging 23 (2022) 402–411

    work page 2022

  2. [2]

    Raimondi, D

    F. Raimondi, D. Martins, R. Coenen, E. Panaioli, D. Khraiche, N. Boddaert, D. Bonnet, M. Atkins, H. El-Said, L. Alshawabkeh, A. Hsiao, Prevalence of Venovenous Shunting and High-Output State Quantified with 4D Flow MRI in Patients with Fontan Circulation, Radiology: Cardiothoracic Imaging 3 (2021) e210161

    work page 2021

  3. [3]

    Sotelo, P

    J. Sotelo, P. Franco, A. Guala, L. Dux-Santoy, A. Ruiz-Muñoz, A. Evangelista, H. Mella, J. Mura, D. E. Hurtado, J. F. Rodríguez-Palomares, S.Uribe, FullyThree-DimensionalHemodynamicCharacterizationofAlteredBloodFlowinBicuspidAorticValvePatientsWithRespectto Aortic Dilatation: A Finite Element Approach, Frontiers in Cardiovascular Medicine 9 (2022) 885338

    work page 2022

  4. [4]

    S.Uribe,P.Beerbaum,T.S.Sørensen,A.Rasmusson,R.Razavi,T.Schaeffter,Four-dimensional(4D)flowofthewholeheartandgreatvessels using real-time respiratory self-gating: Real-Time Self-Gating 4D Flow Imaging, Magnetic Resonance in Medicine 62 (2009) 984–992

    work page 2009

  5. [5]

    M. M. Bissell, F. Raimondi, L. Ait Ali, B. D. Allen, A. J. Barker, A. Bolger, N. Burris, C.-J. Carhäll, J. D. Collins, T. Ebbers, C. J. Francois, A.Frydrychowicz,P.Garg,J.Geiger,H.Ha,A.Hennemuth,M.D.Hope,A.Hsiao,K.Johnson,S.Kozerke,L.E.Ma,M.Markl,D.Martins, M. Messina, T. H. Oechtering, P. Van Ooij, C. Rigsby, J. Rodriguez-Palomares, A. A. W. Roest, A. Ro...

    work page 2023

  6. [6]

    Zhong, E

    L. Zhong, E. M. Schrauben, J. Garcia, S. Uribe, S. M. Grieve, M. S. Elbaz, A. J. Barker, J. Geiger, S. Nordmeyer, A. Marsden, M. Carlsson, R.Tan,P.Garg,J.J.Westenberg,M.Markl,T.Ebbers, Intracardiac4DFlowMRIinCongenitalHeartDisease:RecommendationsonBehalf of the ISMRM Flow & Motion Study Group, Journal of Magnetic Resonance Imaging 50 (2019) 677–681

    work page 2019

  7. [7]

    Isorni, L

    M.-A. Isorni, L. Moisson, N. B. Moussa, S. Monnot, F. Raimondi, R. Roussin, A. Boet, I. Van Aerschot, E. Fournier, S. Cohen, M. Kara, S. Hascoet, 4D flow cardiac magnetic resonance in children and adults with congenital heart disease: Clinical experience in a high volume center, International Journal of Cardiology 320 (2020) 168–177

    work page 2020

  8. [8]

    M.Le,J.Lieman-Sifry,F.Lau,S.Sall,A.Hsiao,D.Golden, Computationallyefficientcardiacviewsprojectionusing3dconvolutionalneural networks, in: International Workshop on Deep Learning in Medical Image Analysis, Springer, pp. 109–116

    work page

Show all 35 references
  1. [9]

    Blansit, T

    K. Blansit, T. Retson, E. Masutani, N. Bahrami, A. Hsiao, Deep Learning–based Prescription of Cardiac MRI Planes, Radiology: Artificial Intelligence 1 (2019) e180069

    2019

  2. [10]

    J.F.Rodríguez-Palomares,L.Dux-Santoy,A.Guala,R.Kale,G.Maldonado,G.Teixidó-Turà,L.Galian,M.Huguet,F.Valente,L.Gutierrez, et al., Aortic flow patterns and wall shear stress maps by 4d-flow cardiovascular magnetic resonance in the assessment of aortic dilatation in bicuspid aorti...

    2018

  3. [11]

    M.Bustamante,S.Petersson,J.Eriksson,U.Alehagen,P.Dyverfeldt,C.-J.Carlhäll,T.Ebbers, Atlas-basedanalysisof4dflowcmr:automated vessel segmentation and flow quantification, Journal of Cardiovascular Magnetic Resonance 17 (2015) 87

    2015

  4. [12]

    :Preprint submitted to Elsevier Page 14 of 16

    X.Fan,Y.Wang,Y.Zhang,M.Bao,B.Jia,D.Lu,Y.Gu,J.Cheng,H.Zhu, Avp-ap:Self-supervisedautomaticviewpositioningin3dcardiac ct via atlas prompting, IEEE Transactions on Medical Imaging 44 (2025). :Preprint submitted to Elsevier Page 14 of 16

    2025

  5. [13]

    P. A. Corrado, D. P. Seiter, O. Wieben, Automatic measurement plane placement for 4D Flow MRI of the great vessels using deep learning, International Journal of Computer Assisted Radiology and Surgery 17 (2022) 199–210

    2022

  6. [14]

    Alansary, L

    A. Alansary, L. L. Folgoc, G. Vaillant, O. Oktay, Y. Li, W. Bai, J. Passerat-Palmbach, R. Guerrero, K. Kamnitsas, B. Hou, S. McDonagh, B. Glocker, B. Kainz, D. Rueckert, Automatic View Planning with Multi-scale Deep Reinforcement Learning Agents, 2018. Accepted by MICCAI 2018

    2018

  7. [15]

    H. Dou, X. Yang, J. Qian, W. Xue, H. Qin, X. Wang, L. Yu, S. Wang, Y. Xiong, P.-A. Heng, D. Ni, Agent with Warm Start and Active Termination for Plane Localization in 3D Ultrasound, 2019. Accepted by MICCAI 2019

    2019

  8. [16]

    Accepted by MICCAI 2020

    Y.Huang,X.Yang,R.Li,J.Qian,X.Huang,W.Shi,H.Dou,C.Chen,Y.Zhang,H.Luo,A.Frangi,Y.Xiong,D.Ni,SearchingCollaborative Agents for Multi-plane Localization in 3D Ultrasound, 2020. Accepted by MICCAI 2020

    2020

  9. [17]

    A.Hennemuth,O.Friman,C.Schumann,J.Bock,J.Drexl,M.Huellebrand,M.Markl,H.-O.Peitgen,Fastinteractiveexplorationof4DMRI flow data, volume 7964, SPIE, pp. 110–120

  10. [18]

    Publisher: Elsevier

    S.M.Pizer,E.P.Amburn,J.D.Austin,R.Cromartie,A.Geselowitz,T.Greer,B.terHaarRomeny,J.B.Zimmerman,K.Zuiderveld, Adaptive histogram equalization and its variations, Computer Vision, Graphics, and Image Processing 39 (1987) 355–368. Publisher: Elsevier

    1987

  11. [19]

    Grondman, L

    I. Grondman, L. Busoniu, G. A. D. Lopes, R. Babuska, A Survey of Actor-Critic Reinforcement Learning: Standard and Natural Policy Gradients, IEEE Transactions on Systems, Man, and Cybernetics, Part C (Applications and Reviews) 42 (2012) 1291–1307

    2012

  12. [20]

    V. Mnih, A. P. Badia, M. Mirza, A. Graves, T. P. Lillicrap, T. Harley, D. Silver, K. Kavukcuoglu, Asynchronous Methods for Deep Reinforcement Learning, 2016. ArXiv:1602.01783 [cs]

    2016 arXiv

  13. [21]

    K.He,X.Zhang,S.Ren,J.Sun, DeepResidualLearningforImageRecognition, in:2016IEEEConferenceonComputerVisionandPattern Recognition (CVPR), IEEE, Las Vegas, NV, USA, 2016, pp. 770–778

    2016

  14. [22]

    Hochreiter, J

    S. Hochreiter, J. Schmidhuber, Long Short-Term Memory, Neural Computation 9 (1997) 1735–1780

    1997

  15. [23]

    D. P. Kingma, J. Ba, Adam: A Method for Stochastic Optimization, 2017. Published as a conference paper at the International Conference for Learning Representations, 2015

    2017

  16. [24]

    Kostrikov, PyTorch Implementations of Asynchronous Advantage Actor Critic, 2018

    I. Kostrikov, PyTorch Implementations of Asynchronous Advantage Actor Critic, 2018. GitHub repository

    2018

  17. [25]

    F. J. Massey Jr, The kolmogorov-smirnov test for goodness of fit, Journal of the American statistical Association 46 (1951) 68–78

    1951

  18. [26]

    J. O. Wobbrock, L. Findlater, D. Gergle, J. J. Higgins, The aligned rank transform for nonparametric factorial analyses using only anova procedures, in: Proceedings of the SIGCHI conference on human factors in computing systems, pp. 143–146

  19. [27]

    L. A. Elkin, M. Kay, J. J. Higgins, J. O. Wobbrock, An aligned rank transform procedure for multifactor contrast tests, in: The 34th annual ACM symposium on user interface software and technology, pp. 754–768

  20. [28]

    T. Chan, L. Vese, Active contours without edges, IEEE Transactions on Image Processing 10 (2001) 266–277

    2001

  21. [29]

    C.Montalba,J.Urbina,J.Sotelo,M.E.Andia,C.Tejos,P.Irarrazaval,D.E.Hurtado,I.Valverde,S.Uribe, Variabilityof4Dflowparameters when subjected to changes in MRI acquisition parameters using a realistic thoracic aortic phantom, Magnetic Resonance in Medicine 79 (2018) 1882–1892

    2018

  22. [30]

    A. A. Taha, A. Hanbury, Metrics for evaluating 3d medical image segmentation: analysis, selection, and tool, BMC medical imaging 15 (2015) 29

    2015

  23. [31]

    Levene, Robust tests for equality of variances, Contributions to probability and statistics (1960) 278–292

    H. Levene, Robust tests for equality of variances, Contributions to probability and statistics (1960) 278–292

    1960

  24. [32]

    M. E. Casciaro, A. F. Pascaner, F. N. Guilenea, J. Alcibar, U. Gencer, G. Soulat, E. Mousseaux, D. Craiem, 4D flow MRI: impact of region of interest size, angulation and spatial resolution on aortic flow assessment, Physiological Measurement 42 (2021) 035004

    2021

  25. [33]

    P.Dyverfeldt,M.Bissell,A.J.Barker,A.F.Bolger,C.-J.Carlhäll,T.Ebbers,C.J.Francois,A.Frydrychowicz,J.Geiger,D.Giese,M.D.Hope, P. J. Kilner, S. Kozerke, S. Myerson, S. Neubauer, O. Wieben, M. Markl, 4D flow cardiovascular magnetic resonance consensus statement, Journal of Cardiov...

    2015

  26. [34]

    T. Gu, F. R. Korosec, W. F. Block, S. B. Fain, Q. Turk, D. Lum, Y. Zhou, T. M. Grist, V. Haughton, C. A. Mistretta, Pc vipr: a high-speed 3d phase-contrast method for flow quantification and high-resolution angiography, American journal of neuroradiology 26 (2005) 743–749

    2005

  27. [35]

    Garcia, K

    J. Garcia, K. Beckie, A. F. Hassanabad, A. Sojoudi, J. A. White, Aortic and mitral flow quantification using dynamic valve tracking and machinelearning:Prospectivestudyassessingstaticanddynamicplanerepeatability,variabilityandagreement, JRSMCardiovascularDisease 10 (2021) 2048...

    2021

Pith tools

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