Pith. sign in

REVIEW 3 major objections 5 minor 47 references

Surrogate Supervision for Robust and Generalizable Deformable Image Registration

T0 review · 3 major / 5 minor · reviewed 2026-08-04 · deepseek-v4-flash

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

Pith's one-line read Surrogate supervision trains registration networks on cleaned stand-ins, giving robustness to artifacts and modality gaps with no extra test-time preprocessing.

desk verdict A solid training-recipe paper: the lung mask-agnostic results are strong and the framing is useful, but the artifact experiment omits the natural test-time-preprocessing baseline, so the headline robustness claim is overextended. read the letter →

arxiv 2509.09869 v1 pith:OEKLEBQ4 submitted 2025-09-11 cs.CV cs.AI

Yihao Liu , Junyu Chen , Lianrui Zuo , Shuwen Wei , Brian D. Boyd , Carmen Andreescu , Olusola Ajilore , Warren D. Taylor
show 2 more authors
Aaron Carass Bennett A. Landman
This is my paper
classification cs.CVcs.AI
keywords surrogatesupervisiondeformableimageregistrationrobustnessartifactmask-agnosticmulti-modalsimilaritylossdeeplearning
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper proposes surrogate supervision, a training paradigm for deformable image registration. The network still receives raw image pairs and predicts a deformation field, but the similarity loss is computed on surrogate images—bias-corrected, masked, or same-subject other-modality versions—that correspond to the same anatomy. This decouples the input domain from the supervision domain, so standard losses like mean squared error or normalized cross-correlation can train a model even when raw inputs are corrupted or multimodal. If correct, any registration architecture can be made artifact-robust, mask-agnostic, and multi-modal simply by choosing a trustworthy surrogate domain, with no added cost at deployment. Experiments across three tasks and three architectures show the expected gains while keeping accuracy on clean data.

What carries the argument

The central mechanism is Eq. (3), the surrogate supervision objective. It modifies the standard registration loss by replacing the raw inputs in the similarity term with surrogates: LSim(eIf, eIm∘phi)+lambda LReg(phi). The deformation field is still estimated from the raw pair; the differentiable warping operator carries gradients from the surrogate-domain loss back into the network. This separates the question of what the network sees from the question of where similarity is well defined, letting a trainer use a domain (preprocessed, masked, or mono-modal) that is reliable even when the raw inputs are heterogeneous.

What would settle it

Train a model on surrogates known to be spatially shifted relative to the input anatomy (e.g., a bias-corrected image that was resampled, or a distortion-corrected diffusion volume) and show that registration accuracy collapses or the deformation field becomes systematically biased; alternatively, measure the correlation between surrogate geometric error and registration error across a test set with known spatial distortions.

Watch

Extended reading notes

Core claim

Surrogate supervision claims that a deformation field predicted from a raw pair can be trained using a loss evaluated entirely in a separate, cleaner domain. Concretely, raw fixed and moving images are fed to the network to predict phi; instead of minimizing LSim(If, Im∘phi), one selects surrogate images (eIf, eIm)—bias-corrected, masked, or paired same-subject scans—and minimizes LSim(eIf, eIm∘phi)+lambda LReg(phi). Because the warp is differentiable, gradients flow back through the surrogate loss to the network. The paper reports that this yields resilience to simulated 3x MRI intensity inhomogeneity, makes lung CT registration indifferent to input masking, and lets T2-to-T1 MR registratio

Load-bearing premise

The load-bearing premise is that the surrogate images inhabit the same coordinate space as the raw inputs, so warping the surrogate is equivalent to warping the underlying anatomy; the paper itself also notes that success depends on the availability and quality of surrogates, shifting the central burden to domain knowledge.

Editorial extensions

If this is right

  • Registration models trained this way can run on raw, unprocessed images at test time, eliminating the need to reproduce the training-time preprocessing pipeline.
  • Standard intensity-based losses (MSE, NCC) become sufficient for artifact-robust and multi-modal registration, weakening the need for specialized similarity measures.
  • Established preprocessing algorithms (bias correction, ROI extraction) and paired data can be repurposed as surrogate generators, embedding their domain knowledge into the trained network.
  • Models become insensitive to whether inputs are pre-masked or not, performing consistently on masked, unmasked, and mixed inputs.
  • Because the framework imposes no constraints on architecture or loss form, it can be layered onto any existing registration network.

Reading between the lines

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

  • Editorial: the same surrogate-supervision recipe should generalize to other dense-warp tasks such as optical flow, slice-to-volume alignment, or atlas construction—any setting with a differentiable warp and a trustworthy surrogate domain.
  • Editorial: the choice of surrogate makes the target invariance explicit and controllable; one could train a single architecture for multiple deployment conditions by swapping surrogate generators, effectively treating surrogate selection as a new hyperparameter.
  • Editorial: augmentation and surrogate supervision are complementary, so combining them could yield models that are simultaneously robust to a wide range of artifacts (input variability) and well-supervised (clean targets). A practical caveat the paper leaves open is that surrogates that alter geometry, such as distortion-corrected diffusion MRI, must be handled with care.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 5 minor

Summary. The paper proposes surrogate supervision, a training framework for deformable registration formalized in Eq. (3): the network estimates a deformation field φ from raw inputs (I_f, I_m), but the similarity loss is evaluated on surrogate images (eI_f, eI_m) that correspond to the original pair, after warping eI_m by φ. This decouples the input domain from the supervision domain, allowing standard similarity losses (MSE, NCC) to be used when the raw inputs are corrupted by artifacts, inconsistently masked, or of different modalities. The framework is instantiated for three applications: artifact-robust brain MR registration using N4-corrected surrogates, mask-agnostic lung CT registration using lung-masked surrogates, and T2-to-T1 multi-modal MR registration using same-subject T1 surrogates. Experiments cover three architectures (VoxelMorph, TransMorph, VFA) and report Dice or TRE improvements over the compared baselines in most settings, with the notable exception that VoxelMorph shows no clear artifact-robustness gain at low inhomogeneity levels.

Significance. If the result holds, the framework is a simple and broadly applicable training principle: rather than designing task-specific similarity losses, one chooses a trustworthy surrogate domain and applies standard losses. The paper has several concrete strengths: it tests three architectures across three distinct tasks, uses a deliberately mismatched test regime in the lung experiment, and the surrogate construction is transparent and does not introduce fitted constants beyond regularization weights inherited from prior work. The inference-time pipeline is unchanged, which is a practical advantage. However, the artifact-robustness experiment is incomplete as evidence for the strongest claim: the w/IC baseline is evaluated in an unnatural deployment mode, and the standard 'correct at test time' baseline is missing. The conceptual contribution is also somewhat incremental relative to prior label-map supervision and Cao et al.'s intra-modality supervision, although the paper does position these as special cases. Overall, the central mechanism is credible, but one load-bearing comparison needs to be repaired before the stated conclusions are fully supported.

major comments (3)
  1. [Section IV-B, Figure 2] The artifact experiment compares Ours (trained on raw inputs with N4-corrected surrogates, tested on raw corrupted inputs) against w/IC (trained on N4-corrected images, tested on raw corrupted inputs) and w/oIC (trained on raw, tested on raw). The w/IC arm is not evaluated in its natural deployment configuration. A user who trained on N4-corrected images would apply N4 correction at test time; without that arm, the experiment only shows that surrogate supervision helps when test-time preprocessing is unavailable. It does not establish the abstract's claim that robustness is achieved 'without increasing complexity' relative to standard preprocessing pipelines. Please add the w/IC model with N4 applied at test time, or explicitly reframe the claim as robustness under missing/incomplete test-time preprocessing.
  2. [Section III-A, Eq. (3)] The formulation does not state sufficient conditions on the surrogates (eI_f, eI_m) relative to (I_f, I_m). It says they 'correspond to' the original pair, but the loss is only meaningful if the surrogates share the coordinate frame of the original images; otherwise eI_m ∘ φ supervises the wrong transformation. The three applications satisfy this by construction (N4 correction, masking, and rigidly aligned same-subject T1), but a precise definition should be added, e.g., 'surrogates are obtained by intensity-only transformations of the input pair, or by paired acquisitions already aligned to the input geometry.' Without this, a user applying the framework to surrogates with different acquisition geometry could silently learn a biased deformation. This is a correctness-risk concern for the general formulation, not for the specific experiments.
  3. [Section IV-B, VoxelMorph results] The text states that for VoxelMorph, surrogate supervision yields 'similar performance' to w/oIC at inhomogeneity levels 1× and 2×, and attributes this to architecture capacity. This is a partial counterexample to the claim that surrogate supervision makes any registration network artifact-robust. The paper should either report the paired test statistics for this comparison, or temper the cross-architecture claim. As written, the evidence supports architecture-dependent benefits, not a general property of the framework.
minor comments (5)
  1. [Section IV-D] The multi-modal experiment does not explicitly describe how the moving surrogate is constructed. It should state that the surrogate is the T1-weighted image of the moving subject, pre-aligned with the T2 input, and that this paired image is available only during training.
  2. [Section IV-B] Typo: 'VoxelMorph, surrogate supervision yield similar performance' should be 'yields.' Also, Figure 2 is dense with significance brackets across three architectures and three inhomogeneity levels; consider separate panels or reporting exact p-values.
  3. [Table II] The NDV values for NCC-based configurations (including Ours) are substantially larger than for CR/MI in VoxelMorph and TransMorph. The text says this is 'due to our choice of λ values,' but λ for NCC was taken from prior work rather than tuned. Please clarify that λ was not re-tuned for Ours/NCC in this comparison.
  4. [References] Reference [22] has formatting errors ('Me. Ren' and 'P. E. Grant'); Reference [3] lists 'L Tian' with inconsistent capitalization. Check the bibliography for consistency.
  5. [Section IV-C] The 'Random' training setting is described as choosing both inputs either masked or unmasked per sample. It may be worth clarifying whether mixed masked/unmasked pairs (one masked, one not) were ever presented during training; this is the setting that appears at test time as 'Mixed.'

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: surrogate supervision is a genuine training-objective modification; missing baselines are experimental-design concerns, not circular derivation.

full rationale

The derivation chain is self-contained. Equation (3) defines the surrogate-supervision objective Ls(ẽ_f, ẽ_m, φ) = LSim(ẽ_f, ẽ_m∘φ) + λLReg(φ), where (ẽ_f, ẽ_m) are independently constructed surrogates (N4-corrected, masked, or paired same-subject images) and φ = N_θ(I_f, I_m) is predicted from the raw inputs. The surrogates are not defined in terms of the prediction, and the loss is not a disguised version of the evaluation metric; robustness must be learned by the network and is not guaranteed (indeed VoxelMorph fails to improve at 1×/2× inhomogeneity). The artifact, mask, and multi-modal experiments compare fixed training configurations under a common test protocol; no fitted parameter is renamed as a prediction, and λ choices are made for baselines or adopted from prior work. Self-citations (e.g., VFA [36], NDV [37], CR [16]) are used as architectures, metrics, or baseline losses, not as load-bearing support for the surrogate-supervision claim. The skeptic's concern—that the artifact experiment omits the natural w/IC + test-time-N4 preprocessing baseline—is a potentially valid threat to the paper's practical conclusion ('without increasing complexity'), but it is an experimental-design / external-validity issue, not a circularity: the claimed training mechanism is not equivalent to its inputs by definition. There is no 'uniqueness theorem' imported from the authors, and no ansatz smuggled in via citation. Therefore no circular step can be exhibited, and the appropriate finding is no significant circularity.

Assumptions & free parameters 2 free parameters · 4 assumptions · 0 invented entities

The paper introduces the concept of surrogate supervision as a training loss design, but no new physical entity, force, or data structure. Its extra cost is only the choice and construction of surrogate images.

free parameters (2)
  • Regularization weight lambda for CR and MI multi-modal baselines = 5
    Tuned on the TransMorph validation set over {0.5, 1, 2, 5, 10}, then fixed for VoxelMorph and VFA; NCC and Ours kept lambda=1 from prior work without comparable tuning, an asymmetric comparison.
  • Lung CT combined loss weights = not stated
    The paper says the combination of NCC, TRE, and diffusion regularizer uses weights adopted from previous work [36] but does not list them, hindering exact reproduction.
assumptions (4)
  • domain assumption Each surrogate is geometrically aligned to its original image, so warping the surrogate equals warping the anatomy.
    Used implicitly in Section III-A and required by all three applications: N4-corrected surrogates in IV-B, masked surrogates in IV-C, and paired T1 surrogates in IV-D.
  • domain assumption Mono-modal similarity (MSE or NCC) in the surrogate domain is a reliable proxy for alignment accuracy in the input domain.
    Equation 3 assumes that high similarity between warped and fixed surrogates implies correct alignment of the original images. This is reasonable when geometry is preserved but is not proven for every artifact, mask, or modality case.
  • domain assumption Paired T1 and T2 scans in IXI are in the same space after N4, skull stripping, and ANTs rigid alignment, so the T1 surrogate of the moving subject correctly represents the moving anatomy.
    Section IV-D; no explicit intra-subject registration is described. Residual T1-T2 misalignment would bias the multi-modal supervision target.
  • standard math The differentiable warping operation passes gradients from the surrogate-domain loss to the network parameters.
    Section III-A relies on this standard auto-differentiation property, citing spatial transformer networks [23].

how reviews work

0 comments
Cite this review

Pith. "Pith review of Surrogate Supervision for Robust and Generalizable Deformable Image Registration." pith.science (2026). https://pith.science/paper/OEKLEBQ4

@misc{pith2026250909869,
  author       = {Pith},
  title        = {Pith review of: Surrogate Supervision for Robust and Generalizable Deformable Image Registration},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/OEKLEBQ4}},
  note         = {Machine review of arXiv:2509.09869}
}
read the original abstract

Objective: Deep learning-based deformable image registration has achieved strong accuracy, but remains sensitive to variations in input image characteristics such as artifacts, field-of-view mismatch, or modality difference. We aim to develop a general training paradigm that improves the robustness and generalizability of registration networks. Methods: We introduce surrogate supervision, which decouples the input domain from the supervision domain by applying estimated spatial transformations to surrogate images. This allows training on heterogeneous inputs while ensuring supervision is computed in domains where similarity is well defined. We evaluate the framework through three representative applications: artifact-robust brain MR registration, mask-agnostic lung CT registration, and multi-modal MR registration. Results: Across tasks, surrogate supervision demonstrated strong resilience to input variations including inhomogeneity field, inconsistent field-of-view, and modality differences, while maintaining high performance on well-curated data. Conclusions: Surrogate supervision provides a principled framework for training robust and generalizable deep learning-based registration models without increasing complexity. Significance: Surrogate supervision offers a practical pathway to more robust and generalizable medical image registration, enabling broader applicability in diverse biomedical imaging scenarios.

Figures

Figures reproduced from arXiv: 2509.09869 by the authors.

Figure 1
Figure 1. Surrogate supervision framework and its applications in label map supervision, multi-modal registration, artifact robustness, and mask agnostic registration. Although the inputs to the registration model may be heterogeneous, we always enforce supervision in a homogeneous domain. to external label maps and train the model by comparing the warped labels to fixed labels. This process is shown in [PITH_FULL_IMAGE:figu… view at source ↗
Figure 2
Figure 2. Quantitative evaluation of registration robustness to MR intensity inhomogeneity. We report the mean Dice Similarity Coefficient (DSC) for each architecture (VoxelMorph, TransMorph, VFA) under three training configurations: with inhomogeneity-correction (w/ IC), without inhomogeneity￾corrected (w/o IC), and surrogate supervision (Ours). Separate box plots are shown for each architecture, with DSC measured at increas… view at source ↗
Figure 3
Figure 3. Examples of test images used to assess artifact robustness. Different degrees of inhomogeneity (up to ×3) were simulated using the N4-estimated inhomogeneity field. The corresponding inhomogeneity-corrected image (×0) is shown for reference. TABLE I QUANTITATIVE RESULTS FOR LUNG CT INHALE-EXHALE REGISTRATION ON THE NLST DATASET. WE REPORT MEAN TARGET REGISTRATION ERROR (TRE) IN MILLIMETERS ON EXPERT-ANNOTATED LANDMA… view at source ↗

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

47 extracted references · 3 linked inside Pith

  1. [1]

    J. Chen, Y . Liu, S. Wei, Z. Bian, S. Subramanian, A. Carass, J. L. Prince, and Y . Du. A survey on deep learning in medical image registration: New technologies, uncertainty, evaluation metrics, and beyond.Medical Image Analysis, 100:103385, 2025

    2025

  2. [2]

    Y . Liu, L. Zuo, S. Han, Y . Xue, J. L. Prince, and A. Carass. Coordinate translator for learning deformable medical image registration. In International Workshop on Multiscale Multimodal Medical Imaging, volume 13594 ofLecture Notes in Computer Science, pages 98–109. Springer, 2022

    2022

  3. [3]

    Vialard, R

    L Tian, H Greer, R Kwitt, F.-X. Vialard, R. San Jos ´e Est ´epar, S. Bouix, R. Rushmore, and M. Niethammer. uniGradICON: A foundation model for medical image registration. In27 th International Conference on Medical Image Computing and Computer Assisted Intervention (MIC- CAI 2024), volume 15002 ofLecture Notes in Computer Science, pages 749–760. Springer, 2024

    2024

  4. [4]

    Balakrishnan, A

    G. Balakrishnan, A. Zhao, M. R. Sabuncu, J. Guttag, and A. V . Dalca. V oxelMorph: A learning framework for deformable medical image registration.IEEE Trans. Med. Imag., 38(8):1788–1800, 2019

    2019

  5. [5]

    Hoffmann, B

    M. Hoffmann, B. Billot, D. N. Greve, J. E. Iglesias, B. Fischl, and A. V . Dalca. SynthMorph: Learning contrast-invariant registration without acquired images.IEEE Trans. Med. Imag., 41(3):543–558, 2021

    2021

  6. [6]

    X. Cao, J. Yang, L. Wang, Z. Xue, Q. Wang, and D. Shen. Deep learning based inter-modality image registration supervised by intra- modality similarity. InMachine Learning in Medical Imaging: 9th International Workshop, MLMI 2018, Held in Conjunction with MICCAI 2018, Granada, Spain, September 16, 2018, Proceedings 9, pages 55–63. Springer, 2018

    2018

  7. [7]

    J. Chen, E. C. Frey, Y . He, W. P. Segars, Y . Li, and Y . Du. TransMorph: Transformer for unsupervised medical image registration.Medical Image Analysis, 82:102615, 2022

    2022

  8. [8]

    Z. Tan, L. Zhang, Y . Lv, Y . Ma, and H. Lu. GroupMorph: Medical image registration via grouping network with contextual fusion.IEEE Trans. Med. Imag., 43(11):3807–3819, 2024

    2024

Show all 47 references
  1. [9]

    J. Chen, S. Wei, J. Honkamaa, P. Marttinen, H. Zhang, M. Liu, Y . Zhou, Z. Tan, Z. Wang, Y . Wang, et al. Beyond the LUMIR challenge: The pathway to foundational registration models.arXiv preprint arXiv:2505.24160, 2025

    2025 arXiv

  2. [10]

    Hering, L

    A. Hering, L. Hansen, T. C. W. Mok, A. C. S. Chung, H. Siebert, S. H¨ager, A. Lange, S. Kuckertz, S. Heldmann, W. Shao, et al. Learn2Reg: Comprehensive multi-task medical image registration challenge, dataset and evaluation in the era of deep learning.IEEE Trans. Med. Imag., 4...

    2023

  3. [11]

    Collignon, F

    A. Collignon, F. Maes, D. Delaere, D. Vandermeulen, P. Suetens, and G. Marchal. Automated multi-modality image registration based on information theory. In14 th Inf. Proc. in Med. Imaging (IPMI 1995), pages 263–274, 1995

    1995

  4. [12]

    W. M. Wells III, P. Viola, H. Atsumi, S. Nakajima, and R. Kikinis. Multi-modal volume registration by maximization of mutual information. Medical Image Analysis, 1(1):35–51, 1996

    1996

  5. [13]

    M. Arar, Y . Ginger, D. Danon, A. H. Bermano, and D. Cohen-Or. Unsupervised multi-modal image registration via geometry preserving image-to-image translation. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 13410–13419, 2020

    2020

  6. [14]

    C. K. Guo. Multi-modal image registration with unsupervised deep learning. M.Eng. Thesis, Massachusetts Institute of Technology, May 2019

    2019

  7. [15]

    Roche, G

    A. Roche, G. Malandain, X. Pennec, and N. Ayache. The correlation ratio as a new similarity measure for multimodal image registration. In 1st International Conference on Medical Image Computing and Computer Assisted Intervention (MICCAI 1998), volume 1496 ofLecture Notes in Co...

    1998

  8. [16]

    X. Chen, Y . Liu, S. Wei, A. Carass, Y . Du, and J. Chen. Correlation ratio for unsupervised learning of multi-modal deformable registration. In O. Colliot and J. Mitra, editors,Medical Imaging 2025: Image Processing, volume 13406, page 1340632. International Society for Optic...

    2025

  9. [17]

    M. P. Heinrich, M. Jenkinson, M. Bhushan, T. Matin, F. V . Gleeson, M. Brady, and J. A. Schnabel. MIND: Modality independent neighbour- hood descriptor for multi-modal deformable registration.Medical Image Analysis, 16(7):1432–1435, 2012

    2012

  10. [18]

    M. Chen, A. Jog, A. Carass, and J. L. Prince. Using image synthesis for multi-channel registration of different image modalities. InProceedings of SPIE Medical Imaging (SPIE-MI 2015), Orlando, FL, February 21 – 26, 2015, volume 9413, pages 462–468, 2015

    2015

  11. [19]

    M. Chen, A. Carass, A. Jog, J. Lee, S. Roy, and J. L. Prince. Cross contrast multi-channel image registration using image synthesis for MR brain images.Medical Image Analysis, 36:2–14, 2017

    2017

  12. [20]

    Z. Li, J. Zhang, T. Ma, T. C. W. Mok, Y .-J. Zhou, Z. Chen, X. Ye, L. Lu, and D. Jin. UniReg: Foundation Model for Controllable Medical Image Registration.arXiv preprint arXiv:2503.12868, 2025

    2025 arXiv

  13. [21]

    J. Chen, S. Wei, Y . Liu, A. Carass, and Y . Du. Pretraining Deformable Image Registration Networks with Random Images. In8 th International Conference on Medical Imaging with Deep Learning (MIDL 2025), 2025

    2025

  14. [22]

    N. Dey, B. Billot, H. E. Wong, C. J. Wang, Me. Ren, P. E. Grant, A. V . Dalca, and P. Golland. Learning general-purpose biomedical volume representations using randomized synthesis.arXiv preprint arXiv:2411.02372, 2024

    2024 arXiv

  15. [23]

    Jaderberg, K

    M. Jaderberg, K. Simonyan, A. Zisserman, et al. Spatial transformer networks.Advances in Neural Information Processing Systems, 28, 2015

    2015

  16. [24]

    T. B. Smith. MRI artifacts and correction strategies.Imaging in Medicine, 2(4):445, 2010

    2010

  17. [25]

    J. F. Barrett and N. Keat. Artifacts in CT: Recognition and avoidance. Radiographics, 24(6):1679–1691, 2004

    2004

  18. [26]

    Le Bihan, C

    D. Le Bihan, C. Poupon, A. Amadon, and F. Lethimonnier. Artifacts and pitfalls in diffusion MRI.Jrnl. of Magnetic Resonance Imaging, 24(3):478–488, 2006

    2006

  19. [27]

    J. C. Lacefield. Physics of Ultrasound. In D. R. Dance, S. Christofides, A. D. A. Maidment, I. D. McLean, and K. H. Ng, editors,Diagnostic LIUet al.: SURROGATE SUPERVISION 9 Radiology Physics: A Handbook for Teachers and Students, chapter 12, pages 291–310. International Atomi...

    2014

  20. [28]

    N. J. Tustison, B. B. Avants, P. A. Cook, Y . Zheng, A. Egan, P. A. Yushkevich, and J. C. Gee. N4ITK: Improved N3 Bias Correction.IEEE Trans. Med. Imag., 29(6):1310–1320, 2010

    2010

  21. [29]

    G. H. Glover and N. J. Pelc. An algorithm for the reduction of metal clip artifacts in CT reconstructions.Med. Phys., 8(6):799–807, 1981

    1981

  22. [30]

    G. Wang, D. L. Snyder, J. A. O’Sullivan, and M. W. Vannier. Iterative deblurring for CT metal artifact reduction.IEEE Trans. Med. Imag., 15(5):657–664, 1996

    1996

  23. [31]

    Jezzard, A

    P. Jezzard, A. S. Barnett, and C. Pierpaoli. Characterization of and correction for eddy current artifacts in echo planar diffusion imaging. Mag. Reson. Med., 39(5):801–812, 1998

    1998

  24. [32]

    F. W. Kremkau and K. J. Taylor. Artifacts in ultrasound imaging.Journal of Ultrasound in Medicine, 5(4):227–237, 1986

    1986

  25. [33]

    B. B. Avants, N. J. Tustison, G. Song, P. A. Cook, A. Klein, and J. C. Gee. A reproducible evaluation of ANTs similarity metric performance in brain image registration.NeuroImage, 54(3):2033–2044, 2011

  26. [34]

    B. B. Avants, N. Tustison, and H. Johnson. Advanced normalization tools (ANTS).Insight j, 2(365):1–35, 2009

    2009

  27. [35]

    Wilcoxon

    F. Wilcoxon. Individual comparisons by ranking methods.Biometrics Bulletin, 1(6):80–83, 1945

    1945

  28. [36]

    Y . Liu, J. Chen, L. Zuo, A. Carass, and J. L. Prince. Vector field attention for deformable image registration.Jrnl. of Medical Imaging, 11(6):064001–064001, 2024

    2024

  29. [37]

    Y . Liu, J. Chen, S. Wei, A. Carass, and J. L. Prince. On finite difference Jacobian computation in deformable image registration.International Journal of Computer Vision, 132:3678–3688, 2024

    2024

  30. [38]

    W. D. Taylor, O. Ajilore, H. T. Karim, M. A. Butters, R. Krafty, B. D. Boyd, L. Banihashemi, S. M. Szymkowicz, C. Ryan, J. Hassenstab, B. A. Landman, and C. Andreescu. Assessing depression recurrence, cognitive burden, and neurobiological homeostasis in late life: Design and r...

    2024

  31. [39]

    Isensee, M

    F. Isensee, M. Schell, I. Pflueger, G. Brugnara D. Bonekamp, U. Neu- berger, A. Wick, H.-P. Schlemmer, S. Heiland, W. Wick, M. Bendszus, K. H. Maier-Hein, and P. Kickingereder. Automated brain extraction of multisequence MRI using artificial neural networks.Human Brain Mapping...

    2019

  32. [40]

    X. Yu, Q. Yang, Y . Zhou, L. Y . Cai, R. Gao, H. H. Lee, T. Li, S. Bao, Z. Xu, T. A. Lasko, et al. UNesT: local spatial representation learning with hierarchical transformer for efficient medical segmentation.Medical Image Analysis, 90:102939, 2023

    2023

  33. [41]

    P. Prorok. The National Cancer Institute Multi-Screening Trial.The Canadian Journal of Oncology, 4:98–101, 1994

    1994

  34. [42]

    IXI Brain Development Dataset

    Biomedical Image Analysis Group, Imperial College London. IXI Brain Development Dataset. https://brain-development. org/ixi-dataset/, 2007

    2007

  35. [43]

    Y . Huo, Z. Xu, Y . Xiong, K. Aboud, P. Parvathaneni, S. Bao, C. Bermudez, S. M. Resnick, L. E. Cutting, and B. A. Landman. 3D whole brain segmentation using spatially localized atlas network tiles.NeuroImage, 194:105–119, 2019

    2019

  36. [44]

    Y . Wang, Y . Kordi, S. Mishra, A. Liu, N. A. Smith, D. Khashabi, and H. Hajishirzi. Self-instruct: Aligning language models with self-generated instructions. InProceedings of the 61st Annual Meeting of the Association for Computational Linguistics, pages 13484–13508, 2023

    2023

  37. [45]

    Applications of a deep learning method for anti-aliasing and super-resolution in mri.Magnetic resonance imaging, 64:132–141, 2019

    Can Zhao, Muhan Shao, Aaron Carass, Hao Li, Blake E Dewey, Lotta M Ellingsen, Jonghye Woo, Michael A Guttman, Ari M Blitz, Maureen Stone, et al. Applications of a deep learning method for anti-aliasing and super-resolution in mri.Magnetic resonance imaging, 64:132–141, 2019

    2019

  38. [46]

    Lung ct harmonization of paired reconstruction kernel images using generative adversarial networks.Medical Physics, 51(8):5510–5523, 2024

    Aravind R Krishnan, Kaiwen Xu, Thomas Z Li, Lucas W Remedios, Kim L Sandler, Fabien Maldonado, and Bennett A Landman. Lung ct harmonization of paired reconstruction kernel images using generative adversarial networks.Medical Physics, 51(8):5510–5523, 2024

    2024

  39. [47]

    HACA3: A unified approach for multi-site MR image harmonization, author=Zuo, Lianrui and Liu, Yihao and Xue, Yuan and Dewey, Blake E and Remedios, Samuel W and Hays, Savannah P and Bilgel, Murat and Mowry, Ellen M and Newsome, Scott D and Calabresi, Peter A and others.Computer...

    2023

Pith tools

Reviewed August 4, 2026 · model on record in the stance chip above.