Fine-Tuning TransMorph with Gradient Correlation for Anatomical Alignment

Kartikay Tehlan; Lukas F\"orner; Thomas Wendler

REVIEW 4 major objections 5 minor 11 references

Fine-Tuning TransMorph with Gradient Correlation for Anatomical Alignment

T0 review · 4 major / 5 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read Fine-tuning TransMorph with FAdam and a gradient-correlation similarity term improves anatomical alignment and reduces non-diffeomorphic volume in unsupervised brain MRI registration.

desk verdict Short challenge report claiming FAdam+GC fine-tuning smooths TransMorph deformations; the NDV gain is real but confounded by a doubled regularization weight and checkpoint selection, so the headline claim isn't identifiable. read the letter →

arxiv 2412.20822 v1 pith:ZIMVEGNU submitted 2024-12-30 eess.IV cs.AIcs.CV

Lukas Förner , Kartikay Tehlan , Thomas Wendler This is my paper · ORCID

classification eess.IVcs.AIcs.CV

keywords deformableimageregistrationTransMorphgradientcorrelationFAdambrainMRIunsupervisednon-diffeomorphicvolumeLearn2RegLUMIR

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

Unsupervised brain MRI registration must balance segmentation accuracy against deformation smoothness. This paper proposes fine-tuning a pre-trained TransMorph model with two modifications: the FAdam optimizer and a gradient-correlation (GC) term added to the similarity loss. On Learn2Reg 2024 LUMIR data, the GC-augmented model matches the baseline in landmark accuracy, improves Dice and 95% Hausdorff distance slightly, and cuts non-diffeomorphic volume from 0.3509% to 0.2690%. The authors read the NDV reduction as evidence of smoother, more anatomically plausible deformations, and a qualitative slice comparison supports better alignment at tissue boundaries with fewer artifacts. The contribution is a targeted fine-tuning recipe that improves deformation quality without retraining from scratch.

What carries the argument

The load-bearing object is the gradient-correlation term $L_{GC}=1-\mathrm{GC}(I_t, I_s \circ \phi)$, where $\mathrm{GC}$ is the average normalized cross-correlation of the $x$-, $y$-, and $z$-gradients of the fixed and warped images. This is added to the local normalized cross-correlation intensity loss as $L_{sim}=L_{IC}+\gamma L_{GC}$ with $\gamma=0.5$, and the total objective becomes $L=L_{sim}+\lambda L_{reg}$ with $\lambda=2$, where $L_{reg}$ is the diffusion regularizer from TransMorph. The other component is FAdam, an Adam variant that uses natural-gradient and Riemannian-geometry corrections intended to improve convergence stability. The GC term carries the smoothness claim: by rewarding aligned image gradients rather than only intensities, it pushes the deformation to keep structural boundaries consistent.

What would settle it

Run the same fine-tuning protocol multiple times with different random seeds, and for each seed select the checkpoint by lowest evaluation loss; if the distributions of NDV for baseline and FAdam+GC overlap substantially, the claimed smoothness improvement is not established.

Watch

Extended reading notes

Core claim

The central claim is that adding gradient correlation to the image similarity loss, together with FAdam-based fine-tuning, produces deformation fields that are substantially smoother than the TransMorph baseline while preserving segmentation and landmark accuracy. Quantitatively, Dice rises from 0.7594 to 0.7614, HdDist95 falls from 3.5074 to 3.4899, and NDV falls from 0.3509% to 0.2690%; TRE changes from 2.4225 mm to 2.4599 mm. The authors attribute the smoothness gain to the GC term, since the FAdam-only model has NDV 0.3549%, close to baseline.

Load-bearing premise

The central claim rests on the assumption that the NDV decrease from 0.3509% to 0.2690% is a real effect of the gradient-correlation loss, not an artifact of the chosen checkpoint or of run-to-run variation; the paper reports no uncertainty or significance testing for NDV, TRE, or HdDist95.

Editorial extensions

If this is right

On the LUMIR evaluation, FAdam+GC achieves Dice 0.7614 and HdDist95 3.4899, slightly better than baseline 0.7594 and 3.5074.
NDV drops from 0.3509% to 0.2690%, indicating that the deformation fields contain fewer non-diffeomorphic, folding voxels.
The improvement is concentrated at tissue boundaries where image gradients are high, which is consistent with the HdDist95 gain.
Landmark accuracy is not sacrificed: TRE is 2.4599 mm for FAdam+GC versus 2.4225 mm for baseline.
FAdam alone does not explain the smoothness effect, as the FAdam-only model has NDV 0.3549%, essentially unchanged from baseline.

Reading between the lines

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

An untested implication is that the gradient-correlation term behaves as an implicit regularizer on deformation topology: aligning image gradients may suppress folding directly, so sweeping $\gamma$ would map a dose-response curve between the GC weight and NDV.
The same loss modification could transfer to other intensity-only registration networks, not only TransMorph; the paper only demonstrates it on the provided baseline.
A practical next step the paper leaves open is multi-seed fine-tuning with statistical comparison, since the magnitude of the NDV gain relative to run-to-run variation is not quantified.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Desk Editor's Note

read the letter

Quick read of the six-pager: the new thing is an empirical result—fine-tuning a pre-trained TransMorph with FAdam and a gradient-correlation term on the LUMIR challenge data gives a modest Dice/HdDist95 bump and an NDV drop from 0.351% to 0.269%. The methods section is clear, the metrics are standard, and the authors are honest that the Dice/TRE gains are tiny. That honesty matters.

The soft spot is exactly where the stress-test lands. Eq. 5 sets λ=2, and the text says this was done to account for the similarity loss being 'about double compared to the baseline.' So the FAdam+GC run is trained with twice the diffusion-regularization weight as the baseline (and, presumably, as the FAdam-only run, though that row's λ isn't stated). The NDV improvement is therefore a regularization-weight/selection artifact candidate, not clean evidence for GC. On top of that, both fine-tuned checkpoints were chosen by 'lowest evaluation loss' at epochs 123 and 170, with no variance or significance reported for NDV, TRE, or HdDist95. The 0.08-percentage-point NDV gap is within the noise envelope as far as the paper shows. A single-slice qualitative figure doesn't close that gap.

That said, I don't think this is a dishonest paper. It reads like a competent challenge report, the kind that usually ships with a clear ablation. The missing piece is exactly one: a λ-matched control (GC at λ=1, or λ=2 without GC) plus confidence intervals. Without that, the central claim is not identifiable from the reported comparison.

Who gets value: readers tracking the Learn2Reg LUMIR leaderboard or experimenting with TransMorph fine-tuning recipes. It won't change how I set up registration experiments, but it's a useful data point if someone wants to chase smoother deformations. I'd send it to review—not because it's a breakthrough, but because the confound is fixable and the result, once cleaned up, is worth having on record. I would not cite it in its current form.

Referee Report

4 major / 5 minor

Summary. The paper proposes a fine-tuning strategy for the TransMorph deformable registration model on the Learn2Reg 2024 LUMIR brain MRI dataset. Two modifications are introduced: replacing the Adam optimizer with FAdam, and adding a gradient correlation (GC) term to the image similarity loss. The authors compare a baseline TransMorph model, a model fine-tuned with FAdam only, and a model fine-tuned with FAdam plus GC. They report Dice, TRE, HdDist95, and NDV scores, finding slight Dice/HdDist95 improvements and a notable reduction in NDV for the FAdam+GC model, and they illustrate qualitative differences on a single slice. The central claim is that the GC term improves anatomical alignment and deformation smoothness.

Significance. If the observed NDV reduction is real and attributable to the GC term, the paper offers a simple, low-cost way to improve deformation plausibility in a strong pretrained registration model, which would be a useful contribution to the Learn2Reg challenge community. The paper is concise and the idea is easy to replicate, and the use of a pretrained baseline is a practical, environmentally motivated choice. The main weakness is that the evidence for the central claim is not statistically grounded: there are no variance or significance measures for NDV, TRE, or HdDist95, and the comparison is confounded by a doubled diffusion-regularization weight and an unspecified checkpoint-selection rule. The qualitative analysis is limited to one slice, and the GC improvement is partly circular, since GC is part of the optimized loss.

major comments (4)

[Section 3.1, Table 1] The central claim that FAdam+GC achieves a notable NDV reduction from 0.3509% to 0.2690% is not supported by any measure of variance or statistical significance. NDV is reported as a single scalar without a standard deviation, and the same is true for TRE and HdDist95. The difference is 0.082 percentage points, which may fall within the noise envelope of the evaluation set. Please provide per-subject distributions for all metrics and a paired significance test (e.g., Wilcoxon signed-rank test) or confidence intervals across the evaluation set for the FAdam+GC versus Baseline and FAdam+GC versus FAdam comparisons. Without this, the 'notable reduction' asserted in the abstract and the conclusion is not established.
[Section 2.2, Eq. (5) and Section 3.1] The NDV improvement is confounded by the change in the diffusion regularization weight lambda. The text states lambda = 2 'to account for the similarity loss being now about double compared to the baseline,' implying the baseline used a smaller lambda (likely 1). Since the diffusion regularizer directly penalizes non-smooth deformations, the lower NDV of FAdam+GC could be caused by the stronger regularization pressure rather than by the GC term. The FAdam-only row does not control for this because its lambda is not stated and its checkpoint selection follows the same loss-based rule. A decisive ablation is required: train the baseline and FAdam-only models with lambda = 2 and the same checkpoint-selection rule, and train FAdam+GC with lambda = 1, so that the effect of GC can be isolated from the regularization weight.
[Section 3.1, checkpoint selection] The checkpoint selection criterion is described only as 'the version with the lowest evaluation loss at 123 and 170 epochs, respectively.' It is not specified which split (training, validation, or test) was used for this loss, nor whether the evaluation loss includes the full objective L = L_sim + lambda * L_reg. If the selection loss includes the lambda-weighted regularizer and the GC term, then the selection procedure itself may favor low-NDV and high-GC solutions, independently of any genuine improvement. Please state the split, the selection metric, and the epoch range searched, and ideally select checkpoints by a metric independent of the training objective (e.g., NDV or Dice on a held-out set).
[Section 3.2, Figure 3] The qualitative evidence is based on a single slice (slice 70) of a single image pair (3457 and 3456). The reported GC values of 0.521 (baseline) and 0.528 (FAdam+GC) are used as supporting evidence, but GC is part of the training objective for FAdam+GC, so improving GC on the training distribution is expected and does not independently validate the method. Please report GC on the full evaluation set or a held-out set, and quantify the artifact reduction beyond the one highlighted slice. A single-slice visual inspection is too anecdotal to support the claim of improved boundary alignment.

minor comments (5)

[Section 3.1, Table 1 header] The table header shows 'HdDist95 ↑', but HdDist95 is a distance measure where lower is better, and the text in Section 3.1 states 'HdDist95 ↓'. Please correct the arrow direction.
[Section 2.2, Eq. (1)] The term 'Normalized Gradient Cross Correlation' is used for GC, but the formula is an average of NCC of gradients. Consider renaming to 'Gradient Correlation' (as in the title) and defining NCC clearly; the current notation may be confused with 'normalized cross-correlation' of the images themselves.
[Figure 2 caption] The caption contains 'cfg. figure 2', which appears to be a typo for 'cf. Figure 2'. Please fix.
[Section 1/2.1, FAdam description] The description of FAdam as incorporating 'principles from natural gradient descent and Riemannian geometry' is vague. Consider adding a brief technical detail or equation from the FAdam paper to make the modification concrete, since the paper is otherwise short.
[Section 3.1, dataset split] The paper states the LUMIR dataset contains over 4,000 scans but does not specify how many were used for fine-tuning versus evaluation, nor whether the evaluation pairs are the same across the three models. Please clarify the exact data split and the number of test cases for each metric.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the main evaluation metrics are independent of the optimized objective; the GC-value comparison is self-referential but not load-bearing.

full rationale

The paper's central claims are empirical comparisons of Dice, HdDist95, TRE, and NDV between a pre-trained TransMorph baseline, an FAdam fine-tuned model, and an FAdam+GC model. These target metrics are not arguments in the optimized loss: Lsim combines LNCC and LGC, and L = Lsim + lambda*Lreg (Eqs. 3-5), while NDV, Dice, TRE, and HdDist95 are evaluated post-hoc using challenge metrics. The only self-referential quantity is the reported GC improvement (0.521 vs 0.528) in Section 3.2, which is expected because GC is directly optimized; however, the paper does not use this comparison as a derived prediction, and the central claims about Dice, HdDist95, and NDV do not reduce to the objective by construction. The checkpoint selection by lowest evaluation loss and the hand-set lambda=2 raise genuine experimental-confound and significance concerns, but these are validity threats rather than circularity: no equation in the paper makes the reported NDV or Dice gains mathematically identical to the fitted loss. No load-bearing self-citation, uniqueness claim, or imported ansatz appears; references to TransMorph, FAdam, and GC are external prior work. Therefore the derivation chain is self-contained at the level claimed, with score 0.

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

The paper introduces no new entities. It relies on two hand-tuned hyperparameters (gamma, lambda), an unspecified checkpoint selection procedure, and domain assumptions about gradient correlation and NDV as proxies for anatomical quality.

free parameters (5)

gamma (GC weight) = 0.5
Chosen by hand in Eq. 4 to 'have about equal values' between LIC and LGC; the central result depends on this balance.
lambda (regularization weight) = 2
Chosen in Eq. 5 to account for the similarity loss being 'about double' compared to baseline; affects the smoothness-similarity trade-off.
initial learning rate = half of baseline learning rate
Set to half the baseline value for fine-tuning; no sensitivity analysis is provided.
training epochs = 200
Both fine-tuned models were trained for 200 epochs; the paper does not justify this duration.
checkpoint selection = epochs 123 and 170
Chosen as the version with 'lowest evaluation loss'; the split for this evaluation loss is unspecified, introducing selection bias.

assumptions (5)

domain assumption Gradient correlation is a valid proxy for anatomical alignment.
Section 1 posits that 'alignment of similar structural gradients supports anatomical coherence', which is a biological plausibility assumption not independently verified in the paper.
domain assumption NDV computed via finite-difference Jacobian correctly measures deformation smoothness.
The paper uses NDV from Liu et al. [6] as the smoothness metric; this assumes the chosen Jacobian discretization is a faithful indicator of anatomical plausibility.
domain assumption The LUMIR challenge evaluation set and metrics are representative.
The paper evaluates on challenge data without describing the split or test set composition; it assumes these metrics generalize.
domain assumption Pre-trained TransMorph weights are a suitable initialization for fine-tuning.
The method relies on the quality and availability of the challenge-provided baseline weights; no analysis of initialization sensitivity is given.
ad hoc to paper The hand-set loss weights gamma=0.5 and lambda=2 produce a balanced objective.
These values are chosen to roughly equalize loss terms, not derived from any optimization or validation; the central results depend on this specific balance.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Fine-Tuning TransMorph with Gradient Correlation for Anatomical Alignment." pith.science (2026). https://pith.science/paper/ZIMVEGNU

@misc{pith2026241220822,
  author       = {Pith},
  title        = {Pith review of: Fine-Tuning TransMorph with Gradient Correlation for Anatomical Alignment},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ZIMVEGNU}},
  note         = {Machine review of arXiv:2412.20822}
}

read the original abstract

Unsupervised deep learning is a promising method in brain MRI registration to reduce the reliance on anatomical labels, while still achieving anatomically accurate transformations. For the Learn2Reg2024 LUMIR challenge, we propose fine-tuning of the pre-trained TransMorph model to improve the convergence stability as well as the deformation smoothness. The former is achieved through the FAdam optimizer, and consistency in structural changes is incorporated through the addition of gradient correlation in the similarity measure, improving anatomical alignment. The results show slight improvements in the Dice and HdDist95 scores, and a notable reduction in the NDV compared to the baseline TransMorph model. These are also confirmed by inspecting the boundaries of the tissue. Our proposed method highlights the effectiveness of including Gradient Correlation to achieve smoother and structurally consistent deformations for interpatient brain MRI registration.

Figures

Figures reproduced from arXiv: 2412.20822 by the authors.

**Figure 1.** Moving, fixed, moved (“Baseline”) and moved (“FAdam+GC”) images. Arrows indicate some example areas where the smoothness of the deformation was improved [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗

**Figure 2.** Displacement fields and grids of “Baseline” and “FAdam+GC” - as RGB image coding XYZ displacements (left) and as grid (right). The arrows indicate some example areas where the smoothness of the deformation was improved (cfg. figure 2) [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗

**Figure 3.** GC values of “Baseline” and “FAdam+GC” as well as the difference in GC and intensity values’. The arrow indicates an example of an area where an artifact created by the baseline was reduced. As can be seen in [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

11 extracted references · 4 canonical work pages

[1]

Medical Image Analysis 82, 102615 (Nov 2022)

Chen, J., Frey, E.C., He, Y., Segars, W.P., Li, Y., Du, Y.: TransMorph : Transformer for unsupervised medical image registration. Medical Image Analysis 82, 102615 (Nov 2022). doi:10.1016/j.media.2022.102615, https://www.sciencedirect.com/science/article/pii/S1361841522002432

arXiv 2022
[2]

doi:10.5281/ZENODO.10991880, https://zenodo.org/doi/10.5281/zenodo.10991880

Dorent, R., Kapur, T., Wells, S., Golby, A., Heyer, W., Chen, J., Liu, Y., Heinrich, M., Walter, A., Lindblad, J., Slodaje, N., Paul-Gilloteaux, P., Hansen, L., Domart, M.C., Collinson, L., Jones, M.: Learn2Reg 2024 (Apr 2024). doi:10.5281/ZENODO.10991880, https://zenodo.org/doi/10.5281/zenodo.10991880

work page doi:10.5281/zenodo.10991880 2024
[3]

NeuroImage 263, 119637 (Nov 2022)

Dufumier, B., Grigis, A., Victor, J., Ambroise, C., Frouin, V., Duchesnay, E.: OpenBHB : a Large - Scale Multi - Site Brain MRI Data -set for Age Prediction and Debiasing . NeuroImage 263, 119637 (Nov 2022). doi:10.1016/j.neuroimage.2022.119637, https://www.sciencedirect.com/science/article/pii/S1053811922007522

work page arXiv 2022
[4]

Hiasa, Y., Otake, Y., Takao, M., Matsuoka, T., Takashima, K., Prince, J.L., Sugano, N., Sato, Y.: Cross-modality image synthesis from unpaired data using CycleGAN : Effects of gradient consistency loss and training data size (Jul 2018), http://arxiv.org/abs/1803.06629, arXiv:1803.06629

arXiv 2018
[5]

Hwang, D.: FAdam : Adam is a natural gradient optimizer using diagonal empirical Fisher information (Sep 2024), http://arxiv.org/abs/2405.12807, arXiv:2405.12807

arXiv 2024
[6]

International Journal of Computer Vision 132(9), 3678--3688 (Sep 2024)

Liu, Y., Chen, J., Wei, S., Carass, A., Prince, J.: On Finite Difference Jacobian Computation in Deformable Image Registration . International Journal of Computer Vision 132(9), 3678--3688 (Sep 2024). doi:10.1007/s11263-024-02047-1, https://doi.org/10.1007/s11263-024-02047-1

work page doi:10.1007/s11263-024-02047-1 2024
[7]

Journal of Cognitive Neuroscience 19(9), 1498--1507 (Sep 2007)

Marcus, D.S., Wang, T.H., Parker, J., Csernansky, J.G., Morris, J.C., Buckner, R.L.: Open Access Series of Imaging Studies ( OASIS ): cross-sectional MRI data in young, middle aged, nondemented, and demented older adults. Journal of Cognitive Neuroscience 19(9), 1498--1507 (Sep 2007). doi:10.1162/jocn.2007.19.9.1498

doi:10.1162/jocn.2007.19.9.1498 2007
[8]

IEEE Transactions on Medical Imaging 17(4), 586--595 (Aug 1998)

Penney, G., Weese, J., Little, J., Desmedt, P., Hill, D., hawkes, D.: A comparison of similarity measures for use in 2- D -3- D medical image registration. IEEE Transactions on Medical Imaging 17(4), 586--595 (Aug 1998). doi:10.1109/42.730403, https://ieeexplore.ieee.org/document/730403/?arnumber=730403, conference Name: IEEE Transactions on Medical Imaging

doi:10.1109/42.730403 1998

Show all 11 references

[9]

Scientific Data 10(1), 449 (Jul 2023)

Taha, A., Gilmore, G., Abbass, M., Kai, J., Kuehn, T., Demarco, J., Gupta, G., Zajner, C., Cao, D., Chevalier, R., Ahmed, A., Hadi, A., Karat, B.G., Stanley, O.W., Park, P.J., Ferko, K.M., Hemachandra, D., Vassallo, R., Jach, M., Thurairajah, A., Wong, S., Tenorio, M.C., Oguns...

2023 doi
[10]

, " * write output.state after.block = add.period write

ENTRY address author booktitle chapter doi edition editor eid howpublished institution journal key month note number organization pages publisher school series title type url volume year label INTEGERS output.state before.all mid.sentence after.sentence after.block FUNCTION in...
[11]

write newline

" write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 gl...

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.

[1] [1]

Medical Image Analysis 82, 102615 (Nov 2022)

Chen, J., Frey, E.C., He, Y., Segars, W.P., Li, Y., Du, Y.: TransMorph : Transformer for unsupervised medical image registration. Medical Image Analysis 82, 102615 (Nov 2022). doi:10.1016/j.media.2022.102615, https://www.sciencedirect.com/science/article/pii/S1361841522002432

arXiv 2022

[2] [2]

doi:10.5281/ZENODO.10991880, https://zenodo.org/doi/10.5281/zenodo.10991880

Dorent, R., Kapur, T., Wells, S., Golby, A., Heyer, W., Chen, J., Liu, Y., Heinrich, M., Walter, A., Lindblad, J., Slodaje, N., Paul-Gilloteaux, P., Hansen, L., Domart, M.C., Collinson, L., Jones, M.: Learn2Reg 2024 (Apr 2024). doi:10.5281/ZENODO.10991880, https://zenodo.org/doi/10.5281/zenodo.10991880

work page doi:10.5281/zenodo.10991880 2024

[3] [3]

NeuroImage 263, 119637 (Nov 2022)

Dufumier, B., Grigis, A., Victor, J., Ambroise, C., Frouin, V., Duchesnay, E.: OpenBHB : a Large - Scale Multi - Site Brain MRI Data -set for Age Prediction and Debiasing . NeuroImage 263, 119637 (Nov 2022). doi:10.1016/j.neuroimage.2022.119637, https://www.sciencedirect.com/science/article/pii/S1053811922007522

work page arXiv 2022

[4] [4]

Hiasa, Y., Otake, Y., Takao, M., Matsuoka, T., Takashima, K., Prince, J.L., Sugano, N., Sato, Y.: Cross-modality image synthesis from unpaired data using CycleGAN : Effects of gradient consistency loss and training data size (Jul 2018), http://arxiv.org/abs/1803.06629, arXiv:1803.06629

arXiv 2018

[5] [5]

Hwang, D.: FAdam : Adam is a natural gradient optimizer using diagonal empirical Fisher information (Sep 2024), http://arxiv.org/abs/2405.12807, arXiv:2405.12807

arXiv 2024

[6] [6]

International Journal of Computer Vision 132(9), 3678--3688 (Sep 2024)

Liu, Y., Chen, J., Wei, S., Carass, A., Prince, J.: On Finite Difference Jacobian Computation in Deformable Image Registration . International Journal of Computer Vision 132(9), 3678--3688 (Sep 2024). doi:10.1007/s11263-024-02047-1, https://doi.org/10.1007/s11263-024-02047-1

work page doi:10.1007/s11263-024-02047-1 2024

[7] [7]

Journal of Cognitive Neuroscience 19(9), 1498--1507 (Sep 2007)

Marcus, D.S., Wang, T.H., Parker, J., Csernansky, J.G., Morris, J.C., Buckner, R.L.: Open Access Series of Imaging Studies ( OASIS ): cross-sectional MRI data in young, middle aged, nondemented, and demented older adults. Journal of Cognitive Neuroscience 19(9), 1498--1507 (Sep 2007). doi:10.1162/jocn.2007.19.9.1498

doi:10.1162/jocn.2007.19.9.1498 2007

[8] [8]

IEEE Transactions on Medical Imaging 17(4), 586--595 (Aug 1998)

Penney, G., Weese, J., Little, J., Desmedt, P., Hill, D., hawkes, D.: A comparison of similarity measures for use in 2- D -3- D medical image registration. IEEE Transactions on Medical Imaging 17(4), 586--595 (Aug 1998). doi:10.1109/42.730403, https://ieeexplore.ieee.org/document/730403/?arnumber=730403, conference Name: IEEE Transactions on Medical Imaging

doi:10.1109/42.730403 1998