Pith. sign in

REVIEW 1 major objections 6 minor 75 references

FetDTIAlign: A Deep Learning Framework for Affine and Deformable Registration of Fetal Brain dMRI

T0 review · 1 major / 6 minor · reviewed 2026-08-09 · deepseek-v4-flash

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

Pith's one-line read This paper presents FetDTIAlign, a deep learning framework that registers fetal brain diffusion MRI across 23 to 36 weeks of gestation, and reports that it outperforms classical optimization-based tools and a generic deep learning…

desk verdict A competent first deep-learning framework for fetal dMRI registration, but the headline Dice superiority is partly trained into the evaluation and needs independent anatomical validation. read the letter →

arxiv 2502.01057 v3 pith:Q4Q6MJXU submitted 2025-02-03 eess.IV cs.AI

Bo Li , Qi Zeng , Simon K. Warfield , Davood Karimi This is my paper · ORCID
classification eess.IVcs.AI
keywords fetalbraindiffusionMRIimageregistrationdeeplearningtensorimagingwhitemattertractsspatialnormalizationdeformable
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

FetDTIAlign is a deep learning method for spatial normalization of fetal brain diffusion MRI, targeting a task where adult-oriented tools fail: noisy, low-resolution, rapidly changing anatomy with few landmarks. The paper aims to show that a learning-based pipeline with separate affine and deformable stages, dual encoders for fractional anisotropy and tensor data, iterative feature-based inference, and explicit tensor reorientation can register fetal brains from 23 to 36 weeks of gestation to an age-matched template. The authors validate this on an internal cohort at three gestational ages and on external data with different acquisition protocols, reporting that FetDTIAlign consistently achieves the highest white matter tract overlap among the compared methods and the best visual alignment. If the claim holds, tract-specific and voxel-wise group analyses of in-utero diffusion MRI become practical, a step that was previously blocked by unreliable spatial correspondence.

What carries the argument

The load-bearing mechanism is a two-stage learned registration pipeline. In the affine stage, separate encoders for the moving and target images produce feature embeddings, a registration head estimates the affine matrix from the mass centers of those embeddings, and recurrent inference composes iteratively refined transforms, choosing the step with the lowest mean squared error on fractional anisotropy and tensor images. In the deformable stage, shared-weight fractional anisotropy and tensor encoders extract four-scale features that are warped and fed to a recurrent decoder built from correlation-aware multi-window multi-layer perceptron blocks, estimating residual deformations from coarse to fine. A tensor-reorientation layer applies the rotation component of each transformation, obtained by polar (singular value) decomposition of the Jacobian, so fiber orientations stay anatomically consistent; finally, the composed affine-plus-deformable field is applied to the original image in a single interpolation to limit smoothing.

What would settle it

Evaluate FetDTIAlign on the external dataset using independently prepared white matter tract segmentations, and compare overlap scores with the classical baselines; if the advantage disappears, the claimed anatomical superiority is an artifact of optimizing the same tract-overlap metric used for evaluation.

Watch

Extended reading notes

Core claim

FetDTIAlign is presented as the first deep learning framework designed specifically for spatial normalization of fetal brain diffusion MRI. It estimates an affine transform and then a residual deformable field using two domain-specific encoders: one processes fractional anisotropy maps and the other the full diffusion tensor, with a recurrent multi-scale decoder and a differentiable tensor-reorientation layer that keeps fiber orientations anatomically consistent. The authors claim that, across gestational ages 23 to 36 weeks and across 60 white matter tracts, FetDTIAlign consistently achieves the highest tract-overlap score among the compared methods in both the affine and deformable stages, produces visually sharper mean fractional anisotropy images, and maintains more physically plausible deformations than a generic deep learning baseline, while also transferring to an external dataset with different acquisition protocols.

Load-bearing premise

The load-bearing premise is that the hand-verified white matter tract maps used for scoring are an impartial measure of anatomical correspondence, even though the learning-based methods were trained to maximize overlap with those same maps while the classical methods were not.

Editorial extensions

If this is right

  • Spatial normalization of fetal diffusion MRI across a wide gestational range becomes a single network inference, making large-cohort voxel-wise and tract-based analyses practical.
  • Because the learned transformation is applied to the original image only once, the pipeline reduces cumulative interpolation smoothing and helps preserve tract boundaries for downstream analysis.
  • The same estimated transformations can be applied to other diffusion-derived maps, such as mean diffusivity, extending the framework beyond fractional anisotropy and tensor comparisons.
  • Cross-protocol generalization to data acquired with different acquisition schemes suggests the method can be applied to multi-shell fetal diffusion data without retraining.
  • If the tract-overlap results are accepted, tract-specific developmental trajectories across 23 to 36 weeks can be built from cross-sectional in-utero data, supporting detection of early deviations in neurodevelopment.

Reading between the lines

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

  • A fairer head-to-head would evaluate tract overlap on independently prepared segmentations from the external dataset, where the training loss was not applied; this is feasible and would break the circularity of training on the same masks used for scoring.
  • The dual fractional-anisotropy-plus-tensor input is partly redundant because fractional anisotropy is derived from the tensor; the paper's observation that adding the tensor reduced image correlation but improved deformation plausibility suggests the two channels play distinct roles, and replacing fractional anisotropy with other scalar maps could target structures such as the cortical plate.
  • The recurrent inference with mean-squared-error-based model selection is effectively a per-subject test-time optimization; whether this stage should be regarded as unsupervised learning or as a hybrid optimizer is a question the paper does not resolve.
  • Since the fetal white matter atlas provides tract masks, an immediate extension is to use the aligned space for atlas-based tract parcellation, enabling automatic tract-specific measurements in new fetuses without manual tractography.
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

1 major / 6 minor

Summary. The paper introduces FetDTIAlign, a deep learning framework for affine and deformable registration of fetal brain diffusion MRI, with a dual-encoder architecture, iterative feature-based inference, and tensor reorientation. The method is validated on internal BCH data across gestational ages 23–36 weeks and on external dHCP data, using Dice overlap of 60 white-matter tract masks, FA cross-correlation, tensor cosine similarity, and deformation Jacobian measures. The authors claim consistent superiority over two classical baselines (FSL, DTI-TK) and one learning baseline (VoxelMorph), and generalizability to the external dataset.

Significance. If the claimed improvements are supported by unbiased evaluation, the work would be a valuable contribution: it addresses a genuinely difficult problem (fetal dMRI registration with low SNR, rapid development, and few landmarks), covers a broad gestational range and a large set of white-matter tracts, includes careful comparison with several baselines on internal data, and the code is publicly released. The extensive visual inspection and the explicit discussion of DTI-TK's role in template construction are also strengths. However, the central benchmark is compromised by overlap between the evaluation metric and the training objective, and the external validation does not include any competing method. These issues must be resolved before the headline claims can be accepted.

major comments (1)
  1. [§5.4] The 'reference similarity score' in §5.4 is computed by translating each subject image by one voxel and computing CC between the translated and original image. This is not a registration baseline; it measures the sensitivity of CC to a small global shift, not the achievable alignment of any method. The statement that deformable registration 'achieved more accurate alignment than this minor misalignment' (p = 6.6 × 10⁻⁴) is therefore not a meaningful comparison against alternative registration approaches. This analysis should be removed or replaced with a comparison to actual competing methods on the same external data.
minor comments (6)
  1. [Table 2 caption] The caption says '** denotes a p-value ≪ 0.01'; it should be 'p < 0.01' rather than the informal 'much less than' notation, and the exact test (paired t-test) should be stated in the caption.
  2. [§3.1] The sentence 'The features ... were aggregated to estimate the twelve affine parameters by aligning the mass centers of the feature embeddings using a least-squares fit' is unclear: aligning mass centers alone determines only three translation parameters, not the twelve parameters of a general affine transform. Please clarify how the least-squares fit yields the full affine matrix.
  3. [§4.4] The list of FSL TBSS configurations includes a custom pipeline (item 5) that combines FSL and DTI-TK; when reporting 'FSL' results in Table 2 and Figures 7, 11–15, the configuration used should be explicitly identified to avoid confusion.
  4. [§4.1] The text says 'Overlapping subjects exist across age groups' and then states the test age groups consist of 'completely non-overlapping subjects'; please reconcile these statements by specifying whether the non-overlap applies within each age group or across the three test age groups.
  5. [§5.4] The paired t-tests report p-values but the number of pairs used in each comparison is not stated; without this, the reader cannot assess the power or the appropriateness of the paired test.
  6. [§1 and §7] The claim of being the 'first deep learning framework' for fetal brain dMRI registration should be qualified against existing learning-based dMRI registration methods such as DDMReg [32] and Segis-Net [24]; if the novelty is specifically the fetal application, the related work should explicitly explain why these prior methods do not apply.

Circularity Check

1 steps flagged · score 4.0 of 10

The deformable-registration Dice result is partially circular: the same white-matter tract masks from [60] are used both as a training loss term (Ltract) for FetDTIAlign and VoxelMorph and as the primary Dice evaluation metric in Table 2.

  1. fitted input called prediction [Section 3.2 (deformable objective LϕD), Section 4.3/Table 2, Section 4.4 (VoxelMorph baseline)]
    "Section 3.2: "the objective function included spatial correspondence of white matter tracts Ltract quantified with the multi-class average Dice coefficient between aligned tract masks within the brain region of the tract atlas ΩA." Section 4.3: "The quality of affine and deformable registration was evaluated using multiple metrics: the Dice coefficient for the warped white matter tract segmentations..." Section 4.4: "we included the Dice coefficient for available white matter tract segmentations, in addition to the original F A similarity measured using normalized cross-correlation...""

    FetDTIAlign's deformable training minimizes LϕD, which contains Ltract: a multi-class Dice term over the [60] tract-atlas masks, and VoxelMorph is also trained with an auxiliary Dice loss on the same masks. The headline evidence of 'superior anatomical correspondence' is the average Dice on these exact [60] masks reported in Table 2. FSL and DTI-TK never optimize any objective involving these masks, so the Dice gap against the classical baselines partly reflects how well each learned model was fit to the evaluation label rather than an independent measure of anatomical accuracy. The affine Dice is less affected because Ltract is added only in the deformable stage, but the deformable Dice, which carries the main claim, is the same functional that was optimized during training.

full rationale

The paper contains no derivation that is equivalent to its inputs in the sense of Eq. X = Eq. Y, and the method is a genuine empirical deep-learning system rather than a construct that is defined by its evaluation. The main circularity concern is a partial one: the deformable registration networks are explicitly optimized on multi-class Dice over the authors' own [60] tract masks, and the same Dice measure is the primary metric used to support the claim of superior anatomical correspondence, especially against classical methods that never saw those masks. That is a fitted-objective-renamed-as-benchmark issue, not a fully circular derivation. Independent content remains: the affine Dice is not part of the affine training loss, the external dHCP evaluation uses a different protocol and shows generalizability, the authors openly disclose that the DTI-TK-built template favors DTI-TK, and additional metrics (FA cross-correlation, NJD, Tenengrad sharpness, visual inspection) go beyond the Dice objective. However, because the central comparative claim relies substantially on a metric that is also a training loss for the learning-based methods, the score is 4 rather than 0-2.

Assumptions & free parameters 5 free parameters · 4 assumptions · 0 invented entities

The empirical claims rest on hand-tuned loss weights, augmentation ranges, a test-time iteration selection, and self-cited reference atlas and tract masks. No new physical or methodological entities are postulated.

free parameters (5)
  • lambda_affine (loss weight for FA and DTI terms) = 10
    Reported as empirically optimized for validation performance in Section 4.5.
  • lambda_deformable (loss weight for FA, DTI, and tract terms) = 100
    Reported as empirically optimized for validation performance in Section 4.5.
  • gamma (deformation smoothness weight) = 100
    Reported as empirically optimized for validation performance in Section 4.5.
  • augmentation ranges for affine training = scale 0.9 to 1.1, rotation -20 to 20 degrees, translation -4 to 4 voxels
    Hand-chosen uniform ranges for affine data augmentation in Section 3.1.
  • number of recurrent inference iterations i = chosen per test image by lowest MSE on target
    Test-time selection rule for the final affine transformation in Section 3.1.
assumptions (4)
  • domain assumption The GA-specific CRL fetal DTI and FA template is a valid reference space for spatial normalization across 23 to 36 weeks.
    The template is used as the target for all registrations; if it is biased or inaccurate, the evaluation is affected for every method. Sections 4.1 and 4.2.
  • domain assumption White matter tract masks from the authors' prior atlas are accurate and complete enough to serve as ground truth for Dice evaluation.
    Dice of tract masks is the primary anatomical correspondence metric; the paper itself notes the upper bound is below 1 because individual scans may miss one side of a bilateral tract. Section 4.3.
  • standard math Tensor reorientation via Jacobian polar decomposition (SVD) correctly preserves diffusion orientation under affine and nonlinear warps.
    This is the standard DTI reorientation method used in both the affine and deformable stages. Section 3.3.
  • domain assumption dHCP data preprocessed by SHARD and resampled to 1.2 mm is directly comparable to BCH data without additional harmonization.
    The external test uses only FA cross-correlation and no tract Dice; remaining protocol and preprocessing differences could affect the comparison. Section 5.4.

how reviews work

0 comments
Cite this review

Pith. "Pith review of FetDTIAlign: A Deep Learning Framework for Affine and Deformable Registration of Fetal Brain dMRI." pith.science (2026). https://pith.science/paper/Q4Q6MJXU

@misc{pith2026250201057,
  author       = {Pith},
  title        = {Pith review of: FetDTIAlign: A Deep Learning Framework for Affine and Deformable Registration of Fetal Brain dMRI},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/Q4Q6MJXU}},
  note         = {Machine review of arXiv:2502.01057}
}
read the original abstract

Diffusion MRI (dMRI) provides unique insights into fetal brain microstructure in utero. Longitudinal and cross-sectional fetal dMRI studies can reveal crucial neurodevelopmental changes but require precise spatial alignment across scans and subjects. This is challenging due to low data quality, rapid brain development, and limited anatomical landmarks. Existing registration methods, designed for high-quality adult data, struggle with these complexities. To address this, we introduce FetDTIAlign, a deep learning approach for fetal brain dMRI registration, enabling accurate affine and deformable alignment. FetDTIAlign features a dual-encoder architecture and iterative feature-based inference, reducing the impact of noise and low resolution. It optimizes network configurations and domain-specific features at each registration stage, enhancing both robustness and accuracy. We validated FetDTIAlign on data from 23 to 36 weeks gestation, covering 60 white matter tracts. It consistently outperformed two classical optimization-based methods and a deep learning pipeline, achieving superior anatomical correspondence. Further validation on external data from the Developing Human Connectome Project confirmed its generalizability across acquisition protocols. Our results demonstrate the feasibility of deep learning for fetal brain dMRI registration, providing a more accurate and reliable alternative to classical techniques. By enabling precise cross-subject and tract-specific analyses, FetDTIAlign supports new discoveries in early brain development.

Figures

Figures reproduced from arXiv: 2502.01057 by the authors.

Figure 1
Figure 1. Illustrated summary of the study. 2 Background 2.1 Diffusion Tensor Imaging (DTI) DTI is one of the local diffusion models that characterizes the diffusion of water molecules in biological tissues, providing insight into the microstructural integrity of white matter in the brain [43]. The core of DTI is the diffusion tensor, a mathematical construct represented as a 3 × 3 symmetric positive-definite matrix,  D, whi… view at source ↗
Figure 2
Figure 2. Affine registration approach in FetDTIAlign. The fetus shown in this figure has a gesta [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Deformable registration approach in FetDTIAlign using an FA encoder (in cyan), a Tensor [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figures from the paper (16 more)
Figure 5
Figure 5. Figure 5: Spatial normalized FA images of the GA23 group using [PITH_FULL_IMAGE:figures/full_fig_p012_5.png]
Figure 6
Figure 6. Figure 6: Spatial normalized FA images of the GA23 group using [PITH_FULL_IMAGE:figures/full_fig_p012_6.png]
Figure 7
Figure 7. Figure 7: Spatial normalized FA images of the GA23 group using [PITH_FULL_IMAGE:figures/full_fig_p013_7.png]
Figure 8
Figure 8. Figure 8: Results of the GA23 group using DTI-TK with the CRL fetal DTI template as the target [PITH_FULL_IMAGE:figures/full_fig_p013_8.png]
Figure 9
Figure 9. Figure 9: Results of the GA23 group using FetDTIAlign with the CRL fetal DTI and FA template as [PITH_FULL_IMAGE:figures/full_fig_p015_9.png]
Figure 10
Figure 10. Figure 10: Original, rotated, and reoriented DTI templates for a 23-week gestational age (GA), using [PITH_FULL_IMAGE:figures/full_fig_p015_10.png]
Figure 11
Figure 11. Figure 11: Visualization of the spatial correspondence of left and right Striato-Parietal tract [PITH_FULL_IMAGE:figures/full_fig_p015_11.png]
Figure 12
Figure 12. Figure 12: Visualization of the spatial correspondence of left and right Inferior Longitudinal Fascicle [PITH_FULL_IMAGE:figures/full_fig_p016_12.png]
Figure 13
Figure 13. Figure 13: Initial FA images and the results of affine registration for the GA25 group using FSL, [PITH_FULL_IMAGE:figures/full_fig_p016_13.png]
Figure 14
Figure 14. Figure 14: Initial FA images and the results of affine registration for the GA35 group using FSL, [PITH_FULL_IMAGE:figures/full_fig_p017_14.png]
Figure 15
Figure 15. Figure 15: Initial FA images and the results of deformable registration for the GA25 group using [PITH_FULL_IMAGE:figures/full_fig_p018_15.png]
Figure 16
Figure 16. Figure 16: Mean FA skeleton (green) and individual FA map projected onto the skeleton (red) using [PITH_FULL_IMAGE:figures/full_fig_p019_16.png]
Figure 17
Figure 17. Figure 17: Mean FA skeleton (green) and individual FA map projected onto the skeleton (red) using [PITH_FULL_IMAGE:figures/full_fig_p020_17.png]
Figure 18
Figure 18. Figure 18: Mean FA skeleton (green) and individual FA map projected onto the skeleton (red) using [PITH_FULL_IMAGE:figures/full_fig_p021_18.png]
Figure 19
Figure 19. Figure 19: Mean FA skeleton (green) across early brain development stages using the deformable [PITH_FULL_IMAGE:figures/full_fig_p021_19.png]
Figure 20
Figure 20. Figure 20: FA cross-correlation between the CRL fetal FA template and the dHCP FA map at [PITH_FULL_IMAGE:figures/full_fig_p022_20.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

75 extracted references · 71 canonical work pages

  1. [60]

    A detailed spatiotemporal atlas of the white matter tracts for the fetal brain

    Camilo Calixto, Matheus Dorigatti Soldatelli, Camilo Jaimes, Lana Pierotich, Simon K Warfield, Ali Gholipour, and Davood Karimi. A detailed spatiotemporal atlas of the white matter tracts for the fetal brain. Proceedings of the National Academy of Sciences, 122(1):e2410341121, 2025

    work page 2025

  2. [1]

    Quantitative magnetic resonance imaging of brain development in premature and mature newborns

    Petra S H¨ uppi, Simon Warfield, Ron Kikinis, Patrick D Barnes, Gary P Zientara, Ferenc A Jolesz, Miles K Tsuji, and Joseph J Volpe. Quantitative magnetic resonance imaging of brain development in premature and mature newborns. Annals of Neurology: Official Journal of the American Neurological Association and the Child Neurology Society , 43(2):224–235, 1998

    work page 1998

  3. [2]

    As- sessment of cortical gyrus and sulcus formation using mr images in normal fetuses

    Seiji Abe, Kenjiro Takagi, Tatsuo Yamamoto, Yoshitaka Okuhata, and Toshinori Kato. As- sessment of cortical gyrus and sulcus formation using mr images in normal fetuses. Prena- tal Diagnosis: Published in Affiliation With the International Society for Prenatal Diagnosis , 23(3):225–231, 2003

    work page 2003

  4. [3]

    Brugger, Gerda Ricken, Gerlinde M

    Christian Mitter, Andr´ as Jakab, Peter C. Brugger, Gerda Ricken, Gerlinde M. Gruber, Dieter Bettelheim, Anke Scharrer, Georg Langs, Johannes A. Hainfellner, Daniela Prayer, and Gregor Kasprian. Validation of in utero tractography of human fetal commissural and internal capsule fibers with histological structure tensor analysis. Frontiers in neuroanatomy,...

    work page 2015

  5. [4]

    Tract-specific group analysis in fetal cohorts using in utero diffusion tensor imaging

    Shadab Khan, Caitlin K Rollins, Cynthia M Ortinau, Onur Afacan, Simon K Warfield, and Ali Gholipour. Tract-specific group analysis in fetal cohorts using in utero diffusion tensor imaging. In International Conference on Medical Image Computing and Computer-Assisted Intervention , pages 28–35. Springer, 2018

    work page 2018

  6. [5]

    Prohl, B

    A.K. Prohl, B. Scherrer, X. Tomas-Fernandez, P.E. Davis, R. Filip-Dhima, S.P. Prabhu, J.M. Peters, E.M. Bebin, D.A. Krueger, H. Northrup, and J.Y. Wu. Early white matter develop- ment is abnormal in tuberous sclerosis complex patients who develop autism spectrum disorder. Journal of neurodevelopmental disorders , 2019

    work page 2019

  7. [6]

    Wilson, M

    S. Wilson, M. Pietsch, L. Cordero-Grande, A.N. Price, J. Hutter, J. Xiao, L. McCabe, M.A. Rutherford, E.J. Hughes, S.J. Counsell, and J.D. Tournier. Development of human white matter pathways in utero over the second and third trimester. Proceedings of the National Academy of Sciences, 2021

    work page 2021

  8. [7]

    Ashburner and K.J

    J. Ashburner and K.J. Friston. Voxel-based morphometry—the methods. Neuroimage, 2000

    work page 2000

Show all 75 references
  1. [8]

    Smith, M

    S.M. Smith, M. Jenkinson, H. Johansen-Berg, D. Rueckert, T.E. Nichols, C.E. Mackay, K.E. Watkins, O. Ciccarelli, M.Z. Cader, P.M. Matthews, and T.E. Behrens. Tract-based spatial statistics: voxelwise analysis of multi-subject diffusion data. Neuroimage, 2006

    2006

  2. [9]

    Smith, M

    S.M. Smith, M. Jenkinson, M.W. Woolrich, C.F. Beckmann, T.E. Behrens, H. Johansen-Berg, P.R. Bannister, M. De Luca, I. Drobnjak, D.E. Flitney, and R.K. Niazy. Advances in functional and structural MR image analysis and implementation as FSL. Neuroimage, 2004

    2004

  3. [10]

    Jones, M.R

    D.K. Jones, M.R. Symms, M. Cercignani, and R.J. Howard. The effect of filter size on VBM analyses of DT-MRI data. Neuroimage, 2005

    2005

  4. [11]

    Smith, G

    S.M. Smith, G. Kindlmann, and S. Jbabdi. In Diffusion MRI, chapter Cross-subject comparison of local diffusion MRI parameters. Academic Press, 2014

    2014

  5. [12]

    A. Zalesky. Moderating registration misalignment in voxelwise comparisons of DTI data: a performance evaluation of skeleton projection. Magnetic resonance imaging , 2011

    2011

  6. [13]

    Bach, F.B

    M. Bach, F.B. Laun, A. Leemans, C.M. Tax, G.J. Biessels, B. Stieltjes, and K.H. Maier-Hein. Methodological considerations on tract-based spatial statistics (TBSS). Neuroimage, 2014

    2014

  7. [14]

    G. Gong, T. Jiang, C. Zhu, Y. Zang, F. Wang, S. Xie, J. Xiao, and X. Guo. Asymmetry analysis of cingulum based on scale-invariant parameterization by diffusion tensor imaging. Human brain mapping, 2005

    2005

  8. [15]

    Corouge, P.T

    I. Corouge, P.T. Fletcher, S. Joshi, S. Gouttard, and G. Gerig. Fiber tract-oriented statistics for quantitative diffusion tensor MRI analysis. Medical image analysis , 2006

    2006

  9. [16]

    O’Donnell, C.F

    L.J. O’Donnell, C.F. Westin, and A.J. Golby. Tract-based morphometry for white matter group analysis. Neuroimage, 2009

    2009

  10. [17]

    Yushkevich, H

    P.A. Yushkevich, H. Zhang, T.J. Simon, and J.C. Gee. Structure-specific statistical mapping of white matter tracts. Neuroimage, 2009

    2009

  11. [18]

    Pecheva, P

    D. Pecheva, P. Yushkevich, D. Batalle, E. Hughes, P. Aljabar, J. Wurie, J.V. Hajnal, A.D. Edwards, D.C. Alexander, S.J. Counsell, and H. Zhang. A tract-specific approach to assessing white matter in preterm infants. Neuroimage, 2017

    2017

  12. [19]

    Nonrigid registration of 3D tensor medical data

    Juan Ruiz-Alzola, C-F Westin, Simon K Warfield, C Alberola, S Maier, and Ron Kikinis. Nonrigid registration of 3D tensor medical data. Medical image analysis , 6(2):143–161, 2002

    2002

  13. [20]

    Ball, S.J

    G. Ball, S.J. Counsell, M. Anjari, N. Merchant, T. Arichi, V. Doria, M.A. Rutherford, A.D. Edwards, D. Rueckert, and J.P. Boardman. An optimised tract-based spatial statistics protocol for neonates: applications to prematurity and chronic lung disease. Neuroimage, 2010

    2010

  14. [21]

    de Groot, M.W

    M. de Groot, M.W. Vernooij, S. Klein, M.A. Ikram, F.M. Vos, S.M. Smith, W.J. Niessen, and J.L. Andersson. Improving alignment in tract-based spatial statistics: evaluation and optimiza- tion of image registration. Neuroimage, 2013

    2013

  15. [22]

    Schwarz, R.I

    C.G. Schwarz, R.I. Reid, J.L. Gunter, M.L. Senjem, S.A. Przybelski, S.M. Zuk, J.L. Whitwell, P. Vemuri, K.A. Josephs, K. Kantarci, and P.M. Thompson. Improved DTI registration allows voxel-based analysis that outperforms tract-based spatial statistics. Neuroimage, 2014

    2014

  16. [23]

    Diffusion tensor image regis- tration using hybrid connectivity and tensor features

    Qian Wang, Pew-Thian Yap, Guorong Wu, and Dinggang Shen. Diffusion tensor image regis- tration using hybrid connectivity and tensor features. Human brain mapping , 35(7):3529–3546, 2014

    2014

  17. [24]

    B. Li, W.J. Niessen, S. Klein, M. de Groot, M.A. Ikram, M.W. Vernooij, and E.E. Bron. Longitudinal diffusion MRI analysis using Segis-Net: a single-step deep-learning framework for simultaneous segmentation and registration. Neuroimage, 2021

    2021

  18. [25]

    Elastic matching of diffusion tensor images

    Daniel C Alexander and James C Gee. Elastic matching of diffusion tensor images. Computer Vision and Image Understanding , 77(2):233–250, 2000

    2000

  19. [26]

    Spatial transforma- tions of diffusion tensor magnetic resonance images

    Daniel C Alexander, Carlo Pierpaoli, Peter J Basser, and James C Gee. Spatial transforma- tions of diffusion tensor magnetic resonance images. IEEE transactions on medical imaging , 20(11):1131–1139, 2001

    2001

  20. [27]

    Zhang, B.B

    H. Zhang, B.B. Avants, P.A. Yushkevich, J.H. Woo, S. Wang, L.F. McCluskey, L.B. Elman, E.R. Melhem, and J.C. Gee. High-dimensional spatial normalization of diffusion tensor images improves the detection of white matter differences: an example study using amyotrophic lateral sc...

    2007

  21. [28]

    Irfanoglu, A

    M.O. Irfanoglu, A. Nayak, J. Jenkins, E.B. Hutchinson, N. Sadeghi, C.P. Thomas, and C. Pier- paoli. DR-TAMAS: diffeomorphic registration for tensor accurate alignment of anatomical struc- tures. Neuroimage, 2016

    2016

  22. [29]

    A continuous staple for scalar, vector, and tensor images: an application to dti analysis

    Olivier Commowick and Simon K Warfield. A continuous staple for scalar, vector, and tensor images: an application to dti analysis. IEEE transactions on medical imaging , 28(6):838–846, 2008

    2008

  23. [30]

    Unbiased white matter atlas construction using diffusion tensor images

    Hui Zhang, Paul A Yushkevich, Daniel Rueckert, and James C Gee. Unbiased white matter atlas construction using diffusion tensor images. In International conference on medical image computing and computer-assisted intervention , pages 211–218. Springer, 2007

    2007

  24. [31]

    Raffelt, J.D

    D. Raffelt, J.D. Tournier, J. Fripp, S. Crozier, A. Connelly, and O. Salvado. Symmetric diffeo- morphic registration of fibre orientation distributions. Neuroimage, 2011

    2011

  25. [32]

    Zhang, W.M

    F. Zhang, W.M. Wells, and L.J. O’Donnell. Deep diffusion MRI registration (DDMReg): a deep learning method for diffusion MRI registration. IEEE transactions on medical imaging , 2021

    2021

  26. [33]

    Commowick, A

    O. Commowick, A. Maarouf, J.C. Ferr´ e, J.P. Ranjeva, G. Edan, and C. Barillot. Diffusion MRI abnormalities detection with orientation distribution functions: A multiple sclerosis longitudinal study. medical image analysis. Medical image analysis , 2015

    2015

  27. [34]

    Karimi, H

    D. Karimi, H. Kebiri, and A. Gholipour. TBSS++: A novel computational method for tract- based spatial statistics. bioRxiv

  28. [35]

    Garyfallidis, O

    E. Garyfallidis, O. Ocegueda, D. Wassermann, and M. Descoteaux. Robust and efficient linear registration of white-matter fascicles in the space of streamlines. NeuroImage, 2015

    2015

  29. [36]

    Garyfallidis, M.A

    E. Garyfallidis, M.A. Cˆ ot´ e, F. Rheault, J. Sidhu, J. Hau, L. Petit, D. Fortin, S. Cunanne, and M. Descoteaux. Recognition of white matter bundles using local and global streamline-based registration and clustering. NeuroImage, 2018

    2018

  30. [37]

    Unbi- ased groupwise registration of white matter tractography

    Lauren J O’Donnell, William M Wells, Alexandra J Golby, and Carl-Fredrik Westin. Unbi- ased groupwise registration of white matter tractography. In Medical Image Computing and Computer-Assisted Intervention–MICCAI 2012: 15th International Conference, Nice, France, October 1-5,...

    2012

  31. [38]

    Taquet, B

    M. Taquet, B. Scherrer, O. Commowick, J.M. Peters, M. Sahin, B. Macq, and S.K. Warfield. A mathematical framework for the registration and analysis of multi-fascicle models for population studies of the brain microstructure. IEEE transactions on medical imaging , 2013

    2013

  32. [39]

    Jbabdi, T.E

    S. Jbabdi, T.E. Behrens, and S.M. Smith. Crossing fibres in tract-based spatial statistics. Neuroimage, 2010

    2010

  33. [40]

    Keihaninejad, N.S

    S. Keihaninejad, N.S. Ryan, I.B. Malone, M. Modat, D. Cash, G.R. Ridgway, H. Zhang, N.C. Fox, and S. Ourselin. The importance of group-wise registration in tract based spatial statistics study of neurodegeneration: a simulation study in Alzheimer’s disease. PloS one , 2012

    2012

  34. [41]

    B. Li, W.J. Niessen, S. Klein, M.A. Ikram, M.W. Vernooij, and E.E. Bron. Learning unbiased group-wise registration (LUGR) and joint segmentation: evaluation on longitudinal diffusion MRI. In Medical Imaging 2021: Image Processing , 2021

    2021

  35. [42]

    Longitudinal reliability of tract-based spatial statistics in diffusion tensor imaging

    Tara Madhyastha, Susan Merillat, Sarah Hirsiger, Ladina Bezzola, Franziskus Liem, Thomas Grabowski, and Lutz J¨ ancke. Longitudinal reliability of tract-based spatial statistics in diffusion tensor imaging. Human brain mapping , 35(9):4544–4555, 2014

    2014

  36. [43]

    Toward a quantitative assessment of diffusion anisotropy

    Carlo Pierpaoli and Peter J Basser. Toward a quantitative assessment of diffusion anisotropy. Magnetic resonance in Medicine , 36(6):893–906, 1996

    1996

  37. [44]

    Estimation of the effective self-diffusion tensor from the nmr spin echo

    Peter J Basser, James Mattiello, and Denis LeBihan. Estimation of the effective self-diffusion tensor from the nmr spin echo. Journal of Magnetic Resonance, Series B , 103(3):247–254, 1994

    1994

  38. [45]

    Non-linear registration, aka spatial normalisation fmrib technical report tr07ja2

    Jesper LR Andersson, Mark Jenkinson, Stephen Smith, et al. Non-linear registration, aka spatial normalisation fmrib technical report tr07ja2. FMRIB Analysis Group of the University of Oxford, 2(1):1–22, 2007

    2007

  39. [46]

    Accurate and robust brain image alignment using boundary- based registration

    Douglas N Greve and Bruce Fischl. Accurate and robust brain image alignment using boundary- based registration. Neuroimage, 48(1):63–72, 2009

    2009

  40. [47]

    Brian C. Hall. Lie Groups, Lie Algebras, and Representations: An Elementary Introduction, Graduate Texts in Mathematics , volume 222 (2nd ed.). Springer, 2015

    2015

  41. [48]

    Zhang, P

    H. Zhang, P. A. Yushkevich, D. C. Alexander, and J. C. Gee. Deformable registration of diffusion tensor MR images with explicit orientation optimization. Medical image analysis , 2006

    2006

  42. [49]

    Detailed anatomic segmentations of a fetal brain diffusion tensor imaging atlas between 23 and 30 weeks of gestation

    Camilo Calixto, Fedel Machado-Rivas, Davood Karimi, Maria C Cortes-Albornoz, Lina M Acosta-Buitrago, Sebastian Gallo-Bernal, Onur Afacan, Simon K Warfield, Ali Gholipour, and Camilo Jaimes. Detailed anatomic segmentations of a fetal brain diffusion tensor imaging atlas between...

    2023

  43. [50]

    Elastix: a toolbox for intensity-based medical image registration

    Stefan Klein, Marius Staring, Keelin Murphy, Max A Viergever, and Josien PW Pluim. Elastix: a toolbox for intensity-based medical image registration. IEEE transactions on medical imaging, 29(1):196–205, 2009

    2009

  44. [51]

    Rectifier nonlinearities improve neural network acoustic models

    Andrew L Maas, Awni Y Hannun, Andrew Y Ng, et al. Rectifier nonlinearities improve neural network acoustic models. In Proc. icml, volume 30, page 3. Atlanta, GA, 2013

    2013

  45. [52]

    Gradient-based learning applied to document recognition

    Yann LeCun, L´ eon Bottou, Yoshua Bengio, and Patrick Haffner. Gradient-based learning applied to document recognition. Proceedings of the IEEE, 86(11):2278–2324, 1998

    1998

  46. [53]

    Anatomy-aware and acquisition-agnostic joint registration with synthmorph

    Malte Hoffmann, Andrew Hoopes, Douglas N Greve, Bruce Fischl, and Adrian V Dalca. Anatomy-aware and acquisition-agnostic joint registration with synthmorph. Imaging Neu- roscience, 2024

    2024

  47. [54]

    Brainmorph: A foundational keypoint model for robust and flexible brain MRI registration

    Alan Q Wang, Rachit Saluja, Heejong Kim, Xinzi He, Adrian Dalca, and Mert R Sabuncu. Brainmorph: A foundational keypoint model for robust and flexible brain MRI registration. arXiv preprint arXiv:2405.14019 , 2024

    2024 arXiv

  48. [55]

    Mathematical methods and algorithms for signal processing

    Todd K Moon and Wynn C Stirling. Mathematical methods and algorithms for signal processing. (No Title) , 2000

    2000

  49. [56]

    Hammer: hierarchical attribute matching mechanism for elastic registration

    Dinggang Shen and Christos Davatzikos. Hammer: hierarchical attribute matching mechanism for elastic registration. IEEE transactions on medical imaging , 21(11):1421–1439, 2002

    2002

  50. [57]

    Correlation-aware coarse-to-fine mlps for deformable medical image registration

    Mingyuan Meng, Dagan Feng, Lei Bi, and Jinman Kim. Correlation-aware coarse-to-fine mlps for deformable medical image registration. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages 9645–9654, 2024

    2024

  51. [58]

    Fetal diffusion mri acquisition and analysis in the developing human connectome project

    Daan Christiaens, Lucilio Cordero-Grande, Anthony N Price, Jana Hutter, Emer Hughes, Ser- ena J Counsell, J-Donald Tournier, and Joseph V Hajnal. Fetal diffusion mri acquisition and analysis in the developing human connectome project. In ISMRM Annual Meeting & Exhibition, page...

    2019

  52. [59]

    Motion-robust diffusion-weighted brain MRI reconstruction through slice-level registration-based motion tracking

    Bahram Marami, Benoit Scherrer, Onur Afacan, Burak Erem, Simon K Warfield, and Ali Gholipour. Motion-robust diffusion-weighted brain MRI reconstruction through slice-level registration-based motion tracking. IEEE transactions on medical imaging , 35(10):2258–2269, 2016

    2016

  53. [61]

    Anatomically constrained tractography of the fetal brain

    Camilo Calixto, Camilo Jaimes, Matheus D Soldatelli, Simon K Warfield, Ali Gholipour, and Davood Karimi. Anatomically constrained tractography of the fetal brain. NeuroImage, 297:120723, 2024

    2024

  54. [62]

    The white matter query language: a novel approach for describ- ing human white matter anatomy

    Demian Wassermann, Nikos Makris, Yogesh Rathi, Martha Shenton, Ron Kikinis, Marek Ku- bicki, and Carl-Fredrik Westin. The white matter query language: a novel approach for describ- ing human white matter anatomy. Brain Structure and Function , 221(9):4705–4721, 2016

    2016

  55. [63]

    Scat- tered slice shard reconstruction for motion correction in multi-shell diffusion mri

    Daan Christiaens, Lucilio Cordero-Grande, Maximilian Pietsch, Jana Hutter, Anthony N Price, Emer J Hughes, Katy Vecchiato, Maria Deprez, A David Edwards, Joseph V Hajnal, et al. Scat- tered slice shard reconstruction for motion correction in multi-shell diffusion mri. Neuroima...

    2021

  56. [64]

    A unifying theoretical and algorithmic framework for least squares methods of estimation in diffusion tensor imaging

    Cheng Guan Koay, Lin-Ching Chang, John D Carew, Carlo Pierpaoli, and Peter J Basser. A unifying theoretical and algorithmic framework for least squares methods of estimation in diffusion tensor imaging. Journal of magnetic resonance , 182(1):115–125, 2006

    2006

  57. [65]

    Itk-snap: An interactive tool for semi- automatic segmentation of multi-modality biomedical images

    Paul A Yushkevich, Yang Gao, and Guido Gerig. Itk-snap: An interactive tool for semi- automatic segmentation of multi-modality biomedical images. In 2016 38th annual international conference of the IEEE engineering in medicine and biology society (EMBC) , pages 3342–3345. IEEE, 2016

    2016

  58. [66]

    Fetal-bet: Brain extraction tool for fetal mri

    Razieh Faghihpirayesh, Davood Karimi, Deniz Erdo˘ gmu¸ s, and Ali Gholipour. Fetal-bet: Brain extraction tool for fetal mri. IEEE Open Journal of Engineering in Medicine and Biology , 2024

    2024

  59. [67]

    Haitch: A framework for distortion and motion correction in fetal multi-shell diffusion-weighted mri

    Haykel Snoussi, Davood Karimi, Onur Afacan, Mustafa Utkur, and Ali Gholipour. Haitch: A framework for distortion and motion correction in fetal multi-shell diffusion-weighted mri. ArXiv, pages arXiv–2406, 2024

    2024

  60. [68]

    Analysis of focus measure operators for shape-from-focus

    Said Pertuz, Domenec Puig, and Miguel Angel Garcia. Analysis of focus measure operators for shape-from-focus. Pattern Recognition, 46(5):1415–1432, 2013

    2013

  61. [69]

    Voxel- morph: a learning framework for deformable medical image registration

    Guha Balakrishnan, Amy Zhao, Mert R Sabuncu, John Guttag, and Adrian V Dalca. Voxel- morph: a learning framework for deformable medical image registration. IEEE transactions on medical imaging, 38(8):1788–1800, 2019

    2019

  62. [70]

    S. Khan, L. Vasung, B. Marami, C.K. Rollins, O. Afacan, C.M. Ortinau, E. Yang, S.K. Warfield, and A. Gholipour. Fetal brain growth portrayed by a spatiotemporal diffusion tensor MRI atlas computed from in utero images. Neuroimage, 2019

    2019

  63. [71]

    Adam: A method for stochastic optimization

    Diederik P Kingma. Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.6980, 2014

    2014 arXiv

  64. [72]

    Repro- ducibility of diffusion MRI-based tractography in the fetal brain.Journal of Magnetic Resonance Imaging, 2024

    Jiaxin Xiao, Cong Sun, Ruike Chen, Zhiyong Zhao, Guangbin Wang, and Dan Wu. Repro- ducibility of diffusion MRI-based tractography in the fetal brain.Journal of Magnetic Resonance Imaging, 2024

    2024

  65. [73]

    Radial organization of developing preterm human cerebral cortex revealed by non-invasive water diffu- sion anisotropy MRI

    Robert C McKinstry, Amit Mathur, Jeffrey H Miller, Alpay Ozcan, Abraham Z Snyder, Geor- gia L Schefft, C Robert Almli, Shelly I Shiran, Thomas E Conturo, and Jeffrey J Neil. Radial organization of developing preterm human cerebral cortex revealed by non-invasive water diffu- s...

    2002

  66. [74]

    Transmorph: Transformer for unsupervised medical image registration

    Junyu Chen, Eric C Frey, Yufan He, William P Segars, Ye Li, and Yong Du. Transmorph: Transformer for unsupervised medical image registration. Medical image analysis , 82:102615, 2022

    2022

  67. [75]

    Affine medical image registration with coarse-to-fine vision transformer

    Tony CW Mok and Albert Chung. Affine medical image registration with coarse-to-fine vision transformer. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 20835–20844, 2022

    2022

Pith tools

Reviewed August 9, 2026 · model on record in the stance chip above.