Pith. sign in

REVIEW 3 major objections 6 minor 55 references

Efficient and robust 3D blind harmonization for large domain gaps

T0 review · 3 major / 6 minor · reviewed 2026-08-16 · deepseek-v4-flash

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

Pith's one-line read Edges alone can bridge large MRI domain gaps in 3D, using only target-domain training data and a correlation-based refinement step.

desk verdict Solid 3D blind harmonization paper with a load-bearing edge-invariance assumption that is acknowledged but not directly tested; deserves review with requests for statistical rigor. read the letter →

arxiv 2505.00133 v1 pith:PFVFDSL5 submitted 2025-04-30 eess.IV cs.CV

Hwihun Jeong , Hayeon Lee , Se Young Chun , Jongho Lee This is my paper · ORCID
classification eess.IVcs.CV
keywords blindharmonizationMRIedge-to-imagemodelrectifiedflow3DmedicalimagingdomaingapCannyedgedetectionhallucinationsuppression
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

BlindHarmonyDiff claims that an MRI scan from an unseen scanner can be harmonized to a target scanner's style using only a subject-level edge map, without ever seeing source-domain data during training. The paper trains a 3D rectified-flow model on target-domain volumes to reconstruct a full image from its Canny edge map, then applies that model to the edges of source-domain scans. A refinement step that raises normalized cross-correlation with the source scan is shown to be indispensable for keeping lesions and other small structures intact. On T1-weighted traveling subjects across four source scanners and one target scanner, the method reports higher PSNR and SSIM and better downstream segmentation and age-prediction performance than prior blind harmonization methods.

What carries the argument

The load-bearing object is the edge-to-image model: a 3D rectified flow (straight-line transport between Gaussian noise and image patches, Eq. 5) conditioned on a Canny edge map patch and normalized spatial coordinates, trained only on target-domain volumes. It operationalizes the edge-invariance identity $D(x_{\mathrm{tar}})=D(x_{\mathrm{src}})=e$ and generates $x_{\mathrm{har}}=\mathrm{E2I}(D(x_{\mathrm{src}}),\theta^*)$. Supporting machinery includes subject-level Canny thresholding that fixes 8% of voxels as edge voxels, multi-stride patch training that downsamples sub-volumes to $64^3$ patches to blend local and global context under memory constraints, and a refinement module that iteratively raises normalized cross-correlation between the harmonized and source images to suppress hallucination and preserve structures.

What would settle it

Compute the Dice coefficient between subject-level 8% Canny edge maps from the same subject scanned on two different scanners; if the average Dice is substantially below 1 in fine-structure or lesion regions, the edge-invariance identity that the pipeline rests on fails for a meaningful share of inputs, and the harmonized image inherits the discrepancy.

Watch

Extended reading notes

Core claim

The central discovery is that a single edge-to-image model trained only on target-domain images, together with a correlation-based refinement step, can harmonize images from multiple unseen scanners, including large domain gaps that break earlier blind methods. Concretely, the paper formulates the same subject scanned in two domains as sharing an edge map: $D(x_{\mathrm{tar}})=D(x_{\mathrm{src}})=e$, so generating an image from the source edges with a flow model trained on target edges yields a target-contrast image. BlindHarmonyDiff implements this with a 3D rectified flow trained on $64^3$ patches with multi-stride sampling to capture global context, and a refinement module that performs gradient ascent on the normalized cross-correlation between the harmonized volume and the source volume. The paper reports that this 3D processing removes inter-slice heterogeneity, that the refinement module is essential for reliable outputs, and that the method outperforms prior blind harmonization on PSNR and SSIM as well as downstream tissue segmentation and age prediction on four scanners.

Load-bearing premise

The load-bearing premise is that the same subject yields the same Canny edge map across scanners once the threshold is set to keep 8% of voxels as edges; if a source scan's edges differ from the target scan's edges, the harmonized image inherits wrong structure and the refinement module can only partially fix it.

Editorial extensions

If this is right

  • Harmonization no longer requires paired or source-domain training data, so a model built once for a target scanner can be applied to any new scanner without retraining.
  • 3D processing removes inter-slice discontinuities that plague slice-wise harmonization, so downstream volumetric analysis sees more homogeneous inputs.
  • The refinement module's correlation step prevents the generator from deleting or inventing small structures such as lesions, which is a precondition for clinical use.
  • For same-contrast multi-scanner harmonization, the reported PSNR, SSIM, and downstream-task results approach supervised methods trained on each source domain, suggesting blind harmonization is a practical substitute when source data are scarce.
  • The same edge-to-image formulation can be pointed at a different target domain by retraining only on that domain, making the method a reusable harmonization building block.

Reading between the lines

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

  • Because the method's success rests on edges surviving across scanners, a natural stress test is to evaluate it on datasets with pathologies where Canny edges are ambiguous; the refinement module may not rescue all such cases.
  • The edge-invariance identity could be tested directly by measuring edge-map agreement across paired traveling-subject scans; if agreement is high only for large structures, the method's robustness may degrade at fine anatomical detail.
  • The same training recipe could extend to other contrasts or modalities, such as CT to MR or T2 to T1, if a contrast-invariant structural representation replaces Canny edges; the paper's own T2-to-T1 result suggests this is currently limited by edge detection.
  • Multi-stride patch training is a general scheme: since it outperforms whole-image training in the paper's 2D ablation, it may benefit other 3D diffusion tasks beyond harmonization.
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.

Referee Report

3 major / 6 minor

Summary. The paper proposes BlindHarmonyDiff, a 3D blind harmonization framework that trains a rectified-flow-based edge-to-image model on target-domain images only. At inference, it detects edges in a source-domain image and generates a harmonized image, followed by a refinement module that maximizes normalized cross-correlation with the source image to reduce hallucination. The authors also introduce multi-stride patch training for efficient 3D training. Experiments on OASIS3 with four source scanners report PSNR/SSIM improvements over prior blind methods and favorable downstream segmentation and age-prediction results.

Significance. If the central edge-invariance assumption holds, the method is a meaningful advance: it provides a 3D, low-cost blind harmonization approach that does not require source-domain training data, with clear ablations (multi-stride ratio, refinement, conditional vs unconditional, 3D vs 2D) and downstream validation. The paper's strengths include systematic ablations and the novel use of edge-to-image generation for harmonization. However, the load-bearing assumption that edge maps are scanner-invariant is not directly measured, and the quantitative claims lack variance analysis.

major comments (3)
  1. [§3.1, Eq. (2); Supplementary S1] The central assumption D(xtar)=D(xsrc)=e is load-bearing but is not directly verified for the four T1-weighted source domains in Tables 1 and 2. The paper itself notes in Supplementary S1 that the 8% edge-voxel threshold 'might occasionally result in inconsistent edge detection and produce poor edge-to-image model output' and in the Discussion that the T2-to-T1 conversion exhibits 'imperfect edge detection' and 'limited' performance. Because the harmonized image is generated solely from D(xsrc), any failure of Eq. (2) propagates into the output, and the refinement module is only a partial safeguard. I request a quantitative evaluation of the assumption: report edge-map agreement (e.g., Dice or Jaccard) between D(xsrc) and D(xtar) on registered traveling subjects per source domain, and show how harmonization fidelity varies with that agreement. Without this, the claim that the method generalizes to unseen source domains remains conditional.
  2. [Tables 1-4, Table S1] All quantitative results are reported as point estimates without error bars, confidence intervals, or statistical tests. The number of image-level test subjects varies from 7 to 30 across source domains (Table S1), so the reported average differences (e.g., Table 1: BlindHarmonyDiff PSNR 23.4 vs. histogram matching 21.8 on Sonata) may not be significant. The downstream results in Table 2 likewise lack per-subject variance. To support the claim of 'superior performance across scanners', please provide per-subject distributions and paired significance tests (e.g., Wilcoxon signed-rank) against the strongest baseline for each metric, and report the number of subjects in each table.
  3. [§3.5, Table 4] The refinement module is an inference-time gradient ascent on the normalized cross-correlation between the harmonized image and the source image. The authors do not report how the step size (0.02) and the number of iterations (6) were chosen, nor how sensitive the results are to them. More importantly, the objective is defined with respect to the source domain, and the paper argues that source and target images are highly correlated in low frequencies; the reported PSNR/SSIM gains in Table 4 show empirical benefit, but a mechanistic analysis (e.g., comparing the frequency spectra of the refined output against target-domain statistics) would clarify why maximizing source NCC improves target fidelity. Without this, the refinement module risks being seen as a source-domain post-processing rather than a target-domain harmonization step.
minor comments (6)
  1. [§3.2] The threshold adjustment procedure is described only informally ('if the ratio ... exceeds the predetermined value, the threshold is decremented'); please specify the iterative algorithm and convergence criterion.
  2. [§3.3] The text says 'rectified flow [29]', but reference [29] is 'Flow matching for generative modeling' (Lipman et al.); the rectified flow method appears in reference [31] (Liu et al.). Please correct the citation.
  3. [Figure 2] The curve has no error bars; please add inter-subject variability. Also clarify the definition of 'multi-stride patch ratio' in the caption.
  4. [Equation (4)] The use of E2I(D(xsrc), θ*) omits the coordinate conditions (i,j,k) that are part of the model input in Eq. (6); please align notation.
  5. [Tables 1 and 2] Add footnotes indicating the number of test subjects per domain and the metric computation mask (brain mask) for reproducibility.
  6. [Figure 3 caption] The caption contains a typo: 'BlindHarmonDiff' should be 'BlindHarmonyDiff'.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the harmonized image is produced by a target-trained edge-to-image model and evaluated on held-out source-domain subjects, so the central claim does not reduce to its inputs.

full rationale

The central derivation chain is explicit and self-contained: the edge-to-image model is trained on target-domain data via Eq. (3), theta* = argmin L(E2I(D(xtar), theta), xtar), and harmonization is then defined by Eq. (4), xhar = E2I(D(xsrc), theta*). The evaluation compares xhar to the paired target-domain image for traveling subjects from four source scanners, with all PSNR/SSIM, segmentation, and age-prediction metrics computed on held-out source-domain subjects. No target-domain harmonization result is used to fit the model; the only fitted quantities are the training weights from target-domain reconstructions, and those are applied to source-domain edge maps rather than to the evaluation targets. The edge-invariance assumption in Eq. (2), D(xtar) = D(xsrc) = e, is indeed load-bearing, but the paper states it as an assumption, attributes it to prior work including the authors' own BlindHarmony [23], and then empirically probes it through the four-scanner harmonization experiments and the T2-to-T1 conversion test, while explicitly acknowledging failure modes in Supplementary S1 and the Discussion. An imperfect or untested assumption is a correctness risk, not circularity. The refinement module is an inference-time optimization with a fixed NCC objective relative to the source image; its benefit is validated by ablations against mutual-information and no-refinement variants, and it is not fitted to the target-domain evaluation metrics. Hyperparameters such as the 8% edge ratio and the 20-80 multi-stride split are selected using target-domain image reconstruction quality, which is standard model selection rather than a prediction of the source-domain harmonization results. The self-citations to BlindHarmony are used for context and for the shared-edge assumption, but the present method's claims are independently tested against external baselines including DeepHarmony, style transfer, SSIMH, and histogram matching. Therefore the derivation does not reduce by construction or by self-citation to its own inputs.

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

The method rests on the edge-invariance assumption (Eq. 2), the low-frequency difference assumption from the authors' prior work, and a set of tuned hyperparameters (edge ratio, multi-stride ratio, refinement settings). There are no invented physical entities. The central empirical claim is largely supported by held-out source-domain evaluation, but the hyperparameter choices and the approximate edge-invariance assumption bound the strength of the conclusions.

free parameters (5)
  • Canny edge ratio = 8% edge voxels
    Chosen empirically in Supplementary S1 because it gives good edge detection for brain images; load-bearing for the consistency assumption in Eq. (2).
  • Multi-stride patch ratio = 20% multi-stride, 80% simple patches
    Selected by PSNR/SSIM sweep on target domain images in Fig. 2; affects reconstruction quality.
  • Refinement module iterations = 6
    Gradient ascent iterations for normalized cross-correlation; no sensitivity analysis is reported.
  • Refinement step size = 0.02
    Gradient ascent step size for the refinement module; no sensitivity analysis is reported.
  • Rectified flow ODE steps = 16
    Number of sampling steps for the rectified flow solver; no ablation is reported.
assumptions (5)
  • domain assumption Same-subject MR images across scanners preserve edges, so subject-level Canny edge detection with an 8% edge threshold yields D(xtar)=D(xsrc)=e (Eq. 2).
    Sec. 3.1, Eq. 2. Central premise of the edge-to-image harmonization; the authors acknowledge in Discussion that edge maps can be heterogeneous for T2-to-T1.
  • domain assumption Differences between source and target domain images are mostly low-frequency, so increasing normalized cross-correlation between source and generated image improves fidelity without forcing target identity.
    Sec. 3.5, based on the authors' prior BlindHarmony [23]. Justifies the refinement module.
  • domain assumption A 3D rectified flow U-Net trained on 64^3 patches can approximate the target image distribution conditional on edge maps and normalized coordinates.
    Sec. 3.3 and S3.1. Assumes network capacity and training recipe suffice; no formal guarantee is provided.
  • domain assumption FSL FLIRT registration aligns traveling-subject source and target images well enough for PSNR/SSIM to be meaningful.
    Sec. 4.4. Registration errors would bias image-level metrics; the paper acknowledges temporal and registration limitations in Discussion.
  • ad hoc to paper Canny edge detection with a fixed edge-voxel ratio of 8% produces consistent edges across domains.
    Supplementary S1. The 8% value is empirically chosen, not derived, and the paper calls it 'somewhat heuristic'.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Efficient and robust 3D blind harmonization for large domain gaps." pith.science (2026). https://pith.science/paper/PFVFDSL5

@misc{pith2026250500133,
  author       = {Pith},
  title        = {Pith review of: Efficient and robust 3D blind harmonization for large domain gaps},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/PFVFDSL5}},
  note         = {Machine review of arXiv:2505.00133}
}
read the original abstract

Blind harmonization has emerged as a promising technique for MR image harmonization to achieve scale-invariant representations, requiring only target domain data (i.e., no source domain data necessary). However, existing methods face limitations such as inter-slice heterogeneity in 3D, moderate image quality, and limited performance for a large domain gap. To address these challenges, we introduce BlindHarmonyDiff, a novel blind 3D harmonization framework that leverages an edge-to-image model tailored specifically to harmonization. Our framework employs a 3D rectified flow trained on target domain images to reconstruct the original image from an edge map, then yielding a harmonized image from the edge of a source domain image. We propose multi-stride patch training for efficient 3D training and a refinement module for robust inference by suppressing hallucination. Extensive experiments demonstrate that BlindHarmonyDiff outperforms prior arts by harmonizing diverse source domain images to the target domain, achieving higher correspondence to the target domain characteristics. Downstream task-based quality assessments such as tissue segmentation and age prediction on diverse MR scanners further confirm the effectiveness of our approach and demonstrate the capability of our robust and generalizable blind harmonization.

Figures

Figures reproduced from arXiv: 2505.00133 by the authors.

Figure 1
Figure 1. (a) Previous blind harmonization methods have limitations such as 2D reconstruction, moderate image quality, and limited [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. PSNR and SSIM across different multi-stride patch ra [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Visual comparison of harmonization results on traveling subjects, demonstrating the effectiveness of BlindHarmonyDiff against [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Tissue segmentation results for the various harmonization methods, with the segmentation network trained on the target domain [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]
Figure 5
Figure 5. Figure 5: Comparison of 3D vs. 2D image processing networks. [PITH_FULL_IMAGE:figures/full_fig_p006_5.png]
Figure 6
Figure 6. Figure 6: Impact of the refinement module on hallucination mit [PITH_FULL_IMAGE:figures/full_fig_p007_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

55 extracted references · 50 canonical work pages

  1. [1]

    Understanding hallucinations in diffu- sion models through mode interpolation

    Sumukh K Aithal, Pratyush Maini, Zachary C Lipton, and J Zico Kolter. Understanding hallucinations in diffu- sion models through mode interpolation. arXiv preprint arXiv:2406.09358, 2024. 8

    arXiv 2024

  2. [2]

    Deep learning for brain mri segmentation: state of the art and future directions

    Zeynettin Akkus, Alfiia Galimzianova, Assaf Hoogi, Daniel L Rubin, and Bradley J Erickson. Deep learning for brain mri segmentation: state of the art and future directions. Journal of Digital Imaging, 30:449–459, 2017. 1

    work page 2017

  3. [3]

    Auto- mated classification of alzheimer’s disease and mild cogni- tive impairment using a single mri and deep neural networks

    Silvia Basaia, Federica Agosta, Luca Wagner, Elisa Canu, Giuseppe Magnani, Roberto Santangelo, Massimo Filippi, Alzheimer’s Disease Neuroimaging Initiative, et al. Auto- mated classification of alzheimer’s disease and mild cogni- tive impairment using a single mri and deep neural networks. NeuroImage: Clinical, 21:101645, 2019. 1

    work page 2019

  4. [4]

    Deep gener- ative medical image harmonization for improving cross-site generalization in deep learning predictors

    Vishnu M Bashyam, Jimit Doshi, Guray Erus, Dhivya Srini- vasan, Ahmed Abdulkadir, Ashish Singh, Mohamad Habes, Yong Fan, Colin L Masters, Paul Maruff, et al. Deep gener- ative medical image harmonization for improving cross-site generalization in deep learning predictors. Journal of Mag- netic Resonance Imaging, 55(3):908–916, 2022. 1

    work page 2022

  5. [5]

    Har- monizing flows: Leveraging normalizing flows for unsuper- vised and source-free mri harmonization

    Farzad Beizaee, Gregory A Lodygensky, Chris L Adamson, Deanne K Thompson, Jeanie LY Cheong, Alicia J Spittle, Peter J Anderson, Christian Desrosiers, and Jose Dolz. Har- monizing flows: Leveraging normalizing flows for unsuper- vised and source-free mri harmonization. Medical Image Analysis, page 103483, 2025. 1, 2

    work page 2025

  6. [6]

    Memory-efficient 3d de- noising diffusion models for medical image processing

    Florentin Bieder, Julia Wolleb, Alicia Durrer, Robin Sand- kuehler, and Philippe C Cattin. Memory-efficient 3d de- noising diffusion models for medical image processing. In MIDL, 2023. 1, 2, 3, 8

    work page 2023

  7. [7]

    Imunity: a generalizable vae-gan solu- tion for multicenter mr image harmonization.Medical Image Analysis, 88:102799, 2023

    Stenzel Cackowski, Emmanuel L Barbier, Michel Dojat, and Thomas Christen. Imunity: a generalizable vae-gan solu- tion for multicenter mr image harmonization.Medical Image Analysis, 88:102799, 2023. 2

    work page 2023

  8. [8]

    A computational approach to edge detection

    John Canny. A computational approach to edge detection. IEEE TPAMI, (6):679–698, 1986. 3

    work page 1986

Show all 55 references
  1. [9]

    Contourdiff: Unpaired image translation with contourguided diffusion models

    Yuwen Chen, Nicholas Konz, Hanxue Gu, Haoyu Dong, Yaqian Chen, Lin Li, Jisoo Lee, and Maciej A Mazurowski. Contourdiff: Unpaired image translation with contourguided diffusion models. arXiv preprint arXiv:2403.10786, 2(3):4,

  2. [10]

    Solving 3d inverse problems us- ing pre-trained 2d diffusion models

    Hyungjin Chung, Dohoon Ryu, Michael T McCann, Marc L Klasky, and Jong Chul Ye. Solving 3d inverse problems us- ing pre-trained 2d diffusion models. In CVPR, 2023. 1

    2023

  3. [11]

    Predicting brain age with deep learn- ing from raw imaging data results in a reliable and heritable biomarker

    James H Cole, Rudra PK Poudel, Dimosthenis Tsagkra- soulis, Matthan W A Caan, Claire Steves, Tim D Spector, and Giovanni Montana. Predicting brain age with deep learn- ing from raw imaging data results in a reliable and heritable biomarker. NeuroImage, 163:115–124, 2017. 6, 3

    2017

  4. [12]

    Deepharmony: A deep learning approach to contrast harmonization across scanner changes

    Blake E Dewey, Can Zhao, Jacob C Reinhold, Aaron Carass, Kathryn C Fitzgerald, Elias S Sotirchos, Shiv Saidha, Jiwon Oh, Dzung L Pham, Peter A Calabresi, et al. Deepharmony: A deep learning approach to contrast harmonization across scanner changes. Magnetic Resonance Imaging ,...

    2019

  5. [13]

    A disentangled latent space for cross-site mri harmonization

    Blake E Dewey, Lianrui Zuo, Aaron Carass, Yufan He, Yihao Liu, Ellen M Mowry, Scott Newsome, Jiwon Oh, Peter A Calabresi, and Jerry L Prince. A disentangled latent space for cross-site mri harmonization. In MICCAI, 2020. 1, 2

    2020

  6. [14]

    Patched denoising diffusion models for high-resolution im- age synthesis

    Zheng Ding, Mengqi Zhang, Jiajun Wu, and Zhuowen Tu. Patched denoising diffusion models for high-resolution im- age synthesis. In ICLR, 2024. 2, 8

    2024

  7. [15]

    Deep learning-based unlearning of dataset bias for mri harmonisation and confound removal

    Nicola K Dinsdale, Mark Jenkinson, and Ana IL Nambu- rete. Deep learning-based unlearning of dataset bias for mri harmonisation and confound removal. NeuroImage, 228: 117689, 2021. 1

    2021

  8. [16]

    Neural spline flows

    Conor Durkan, Artur Bekasov, Iain Murray, and George Pa- pamakarios. Neural spline flows. NeurIPS, 32, 2019. 2

    2019

  9. [17]

    Harmonization of multi-site diffusion tensor imaging data

    Jean-Philippe Fortin, Drew Parker, Birkan Tunc ¸, Takanori Watanabe, Mark A Elliott, Kosha Ruparel, David R Roalf, Theodore D Satterthwaite, Ruben C Gur, Raquel E Gur, et al. Harmonization of multi-site diffusion tensor imaging data. NeuroImage, 161:149–170, 2017. 1, 2

    2017

  10. [18]

    Deep generative models for 3d medical image synthesis

    Paul Friedrich, Yannik Frisch, and Philippe C Cattin. Deep generative models for 3d medical image synthesis. arXiv preprint arXiv:2410.17664, 2024. 1

    2024 arXiv

  11. [19]

    Multi-site mri harmonization via attention-guided deep domain adaptation for brain dis- order identification

    Hao Guan, Yunbi Liu, Erkun Yang, Pew-Thian Yap, Ding- gang Shen, and Mingxia Liu. Multi-site mri harmonization via attention-guided deep domain adaptation for brain dis- order identification. Medical Image Analysis , 71:102076,

  12. [20]

    Fast image-level mri harmonization via spec- trum analysis

    Hao Guan, Siyuan Liu, Weili Lin, Pew-Thian Yap, and Mingxia Liu. Fast image-level mri harmonization via spec- trum analysis. In MLMI, 2022. 4, 5, 6, 2, 3

    2022

  13. [21]

    Learning a variational network for reconstruction of accelerated mri data

    Kerstin Hammernik, Teresa Klatzer, Erich Kobler, Michael P Recht, Daniel K Sodickson, Thomas Pock, and Florian Knoll. Learning a variational network for reconstruction of accelerated mri data. Magnetic Resonance in Medicine , 79 (6):3055–3071, 2018. 1

    2018

  14. [22]

    Improved optimization for the robust and accurate linear registration and motion correction of brain images

    Mark Jenkinson, Peter Bannister, Michael Brady, and Stephen Smith. Improved optimization for the robust and accurate linear registration and motion correction of brain images. NeuroImage, 17(2):825–841, 2002. 4

    2002

  15. [23]

    Blindharmony:” blind” harmonization for mr images via flow model

    Hwihun Jeong, Heejoon Byun, Dong Un Kang, and Jongho Lee. Blindharmony:” blind” harmonization for mr images via flow model. In ICCV, 2023. 1, 2, 4, 5, 6, 3

    2023

  16. [24]

    Denoising diffusion probabilistic models for 3d medical image generation

    Firas Khader, Gustav M ¨uller-Franzes, Soroosh Tayebi Arasteh, Tianyu Han, Christoph Haarburger, Maximilian Schulze-Hagen, Philipp Schad, Sandy Engel- hardt, Bettina Baeßler, Sebastian Foersch, et al. Denoising diffusion probabilistic models for 3d medical image generation. Sc...

    2023

  17. [25]

    Tackling structural hallucination in im- age translation with local diffusion

    Seunghoi Kim, Chen Jin, Tom Diethe, Matteo Figini, Henry FJ Tregidgo, Asher Mullokandov, Philip Teare, and Daniel C Alexander. Tackling structural hallucination in im- age translation with local diffusion. In ECCV, 2024. 8

    2024

  18. [26]

    Residual and plain convolutional neural networks for 3d brain mri classification

    Sergey Korolev, Amir Safiullin, Mikhail Belyaev, and Yulia Dodonova. Residual and plain convolutional neural networks for 3d brain mri classification. In ISBI. IEEE, 2017. 1

    2017

  19. [27]

    Oasis-3: longitudinal neuroimaging, clin- ical, and cognitive dataset for normal aging and alzheimer disease

    Pamela J LaMontagne, Tammie LS Benzinger, John C Mor- ris, Sarah Keefe, Russ Hornbeck, Chengjie Xiong, Eliza- beth Grant, Jason Hassenstab, Krista Moulder, Andrei G Vlassenko, et al. Oasis-3: longitudinal neuroimaging, clin- ical, and cognitive dataset for normal aging and alz...

    2019

  20. [28]

    Improving 3d imaging with pre-trained perpendicular 2d diffusion models

    Suhyeon Lee, Hyungjin Chung, Minyoung Park, Jonghyuk Park, Wi-Sun Ryu, and Jong Chul Ye. Improving 3d imaging with pre-trained perpendicular 2d diffusion models. InICCV,

  21. [29]

    Flow matching for generative modeling

    Yaron Lipman, Ricky TQ Chen, Heli Ben-Hamu, Maximil- ian Nickel, and Matthew Le. Flow matching for generative modeling. In ICLR, 2023. 1, 3

    2023

  22. [30]

    Style transfer using generative adversarial networks for multi-site mri harmonization

    Mengting Liu, Piyush Maiti, Sophia Thomopoulos, Alyssa Zhu, Yaqiong Chai, Hosung Kim, and Neda Jahanshad. Style transfer using generative adversarial networks for multi-site mri harmonization. In MICCAI, 2021. 1, 2, 4, 5, 6

    2021

  23. [31]

    Flow straight and fast: Learning to generate and transfer data with rectified flow

    Xingchao Liu, Chengyue Gong, and Qiang Liu. Flow straight and fast: Learning to generate and transfer data with rectified flow. In ICLR, 2023. 1

    2023

  24. [32]

    Inter-site and inter-scanner diffusion mri data harmo- nization

    Hengameh Mirzaalian, Lipeng Ning, Peter Savadjiev, Ofer Pasternak, Sylvain Bouix, O Michailovich, Gerald Grant, Christine E Marx, Rajendra A Morey, Laura A Flashman, et al. Inter-site and inter-scanner diffusion mri data harmo- nization. NeuroImage, 135:311–323, 2016. 1

    2016

  25. [33]

    Mri image harmonization using cycle- consistent generative adversarial network

    Gourav Modanwal, Adithya Vellal, Mateusz Buda, and Ma- ciej A Mazurowski. Mri image harmonization using cycle- consistent generative adversarial network. In Medical Imag- ing: Computer-Aided Diagnosis , pages 259–264. SPIE,

  26. [34]

    Understanding ssim

    Jim Nilsson and Tomas Akenine-M ¨oller. Understanding ssim. arXiv preprint arXiv:2006.13846, 2020. 8

    2006 arXiv

  27. [35]

    On standardizing the mr image intensity scale

    L ´aszl´o G Ny´ul and Jayaram K Udupa. On standardizing the mr image intensity scale. Magnetic Resonance in Medicine, 42(6):1072–1081, 1999. 1, 2

    1999

  28. [36]

    New variants of a method of mri scale standardization.IEEE TMI, 19(2):143–150, 2000

    L ´aszl´o G Ny ´ul, Jayaram K Udupa, and Xuan Zhang. New variants of a method of mri scale standardization.IEEE TMI, 19(2):143–150, 2000. 2

    2000

  29. [37]

    On estimation of a probability density func- tion and mode

    Emanuel Parzen. On estimation of a probability density func- tion and mode. The Annals of Mathematical Statistics, 33(3): 1065–1076, 1962. 7

    1962

  30. [38]

    Harmonization of large mri datasets for the anal- ysis of brain imaging patterns throughout the lifespan

    Raymond Pomponio, Guray Erus, Mohamad Habes, Jimit Doshi, Dhivya Srinivasan, Elizabeth Mamourian, Vishnu Bashyam, Ilya M Nasrallah, Theodore D Satterthwaite, Yong Fan, et al. Harmonization of large mri datasets for the anal- ysis of brain imaging patterns throughout the lifesp...

    2020

  31. [39]

    U-net: Convolutional networks for biomedical image segmentation

    Olaf Ronneberger, Philipp Fischer, and Thomas Brox. U-net: Convolutional networks for biomedical image segmentation. In MICCAI, 2015. 4, 6, 1, 2

    2015

  32. [40]

    Progressive distillation for fast sampling of diffusion models

    Tim Salimans and Jonathan Ho. Progressive distillation for fast sampling of diffusion models. In ICLR, 2022. 1

    2022

  33. [41]

    Machine learning on brain mri data for differential diagnosis of parkinson’s disease and pro- gressive supranuclear palsy

    Christian Salvatore, Antonio Cerasa, Isabella Castiglioni, F Gallivanone, A Augimeri, M Lopez, G Arabia, M Morelli, MC Gilardi, and A Quattrone. Machine learning on brain mri data for differential diagnosis of parkinson’s disease and pro- gressive supranuclear palsy. Journal o...

    2014

  34. [42]

    Statistical normalization techniques for magnetic resonance imaging

    Russell T Shinohara, Elizabeth M Sweeney, Jeff Goldsmith, Navid Shiee, Farrah J Mateen, Peter A Calabresi, Sam- son Jarso, Dzung L Pham, Daniel S Reich, Ciprian M Crainiceanu, et al. Statistical normalization techniques for magnetic resonance imaging. NeuroImage: Clinical , 6:...

    2014

  35. [43]

    V olumet- ric analysis from a harmonized multisite brain mri study of a single subject with multiple sclerosis

    Russell T Shinohara, Jiwon Oh, Govind Nair, Peter A Cal- abresi, Christos Davatzikos, Jimit Doshi, Roland G Henry, Gloria Kim, Kristin A Linn, Nico Papinutto, et al. V olumet- ric analysis from a harmonized multisite brain mri study of a single subject with multiple sclerosis....

    2017

  36. [44]

    Fast robust automated brain extraction

    Stephen M Smith. Fast robust automated brain extraction. Human Brain Mapping, 17(3):143–155, 2002. 5

    2002

  37. [45]

    Diffusionblend: Learning 3d image prior through position-aware diffusion score blending for 3d com- puted tomography reconstruction

    Bowen Song, Jason Hu, Zhaoxu Luo, Jeffrey A Fessler, and Liyue Shen. Diffusionblend: Learning 3d image prior through position-aware diffusion score blending for 3d com- puted tomography reconstruction. In ICML, 2024. 1

    2024

  38. [46]

    Consistency models

    Yang Song, Prafulla Dhariwal, Mark Chen, and Ilya Sutskever. Consistency models. In ICML, 2023. 1

    2023

  39. [47]

    Mispel: A super- vised deep learning harmonization method for multi-scanner neuroimaging data

    Mahbaneh Eshaghzadeh Torbati, Davneet S Minhas, Charles M Laymon, Pauline Maillard, James D Wilson, Chang-Le Chen, Ciprian M Crainiceanu, Charles S DeCarli, Seong Jae Hwang, and Dana L Tudorascu. Mispel: A super- vised deep learning harmonization method for multi-scanner neuro...

  40. [48]

    Cannygan: Edge-preserving image trans- lation with disentangled features

    Tianren Wang, Teng Zhang, Liangchen Liu, Arnold Wiliem, and Brian Lovell. Cannygan: Edge-preserving image trans- lation with disentangled features. In ICIP, 2019. 2

    2019

  41. [49]

    Patch diffusion: Faster and more data- efficient training of diffusion models

    Zhendong Wang, Yifan Jiang, Huangjie Zheng, Peihao Wang, Pengcheng He, Zhangyang Wang, Weizhu Chen, Mingyuan Zhou, et al. Patch diffusion: Faster and more data- efficient training of diffusion models. NeurIPS, 36, 2023. 2, 3, 8

    2023

  42. [50]

    Quantitative sus- ceptibility mapping using deep neural network: Qsmnet

    Jaeyeon Yoon, Enhao Gong, Itthi Chatnuntawech, Berkin Bilgic, Jingu Lee, Woojin Jung, Jingyu Ko, Hosan Jung, Kawin Setsompop, Greg Zaharchuk, et al. Quantitative sus- ceptibility mapping using deep neural network: Qsmnet. NeuroImage, 179:199–206, 2018. 1, 8

    2018

  43. [51]

    Reliable source approximation: Source-free unsupervised domain adaptation for vestibular schwannoma mri segmentation

    Hongye Zeng, Ke Zou, Zhihao Chen, Rui Zheng, and Huazhu Fu. Reliable source approximation: Source-free unsupervised domain adaptation for vestibular schwannoma mri segmentation. In MICCAI, 2024. 2

    2024

  44. [52]

    Seg- mentation of brain mr images through a hidden markov ran- dom field model and the expectation-maximization algo- rithm

    Yongyue Zhang, Michael Brady, and Stephen Smith. Seg- mentation of brain mr images through a hidden markov ran- dom field model and the expectation-maximization algo- rithm. IEEE TMI, 20(1):45–57, 2001. 6

    2001

  45. [53]

    A deep learning model inte- grating fcnns and crfs for brain tumor segmentation.Medical Image Analysis, 43:98–111, 2018

    Xiaomei Zhao, Yihong Wu, Guidong Song, Zhenye Li, Yazhuo Zhang, and Yong Fan. A deep learning model inte- grating fcnns and crfs for brain tumor segmentation.Medical Image Analysis, 43:98–111, 2018. 1

    2018

  46. [54]

    Unpaired image-to-image translation using cycle- consistent adversarial networks

    Jun-Yan Zhu, Taesung Park, Phillip Isola, and Alexei A Efros. Unpaired image-to-image translation using cycle- consistent adversarial networks. In ICCV, pages 2223–2232,

  47. [55]

    Unsupervised mr harmonization by learning disentangled representations using information bottleneck theory

    Lianrui Zuo, Blake E Dewey, Yihao Liu, Yufan He, Scott D Newsome, Ellen M Mowry, Susan M Resnick, Jerry L Prince, and Aaron Carass. Unsupervised mr harmonization by learning disentangled representations using information bottleneck theory. NeuroImage, 243:118569, 2021. 2 Effic...

    2021

Pith tools

Reviewed August 16, 2026 · model on record in the stance chip above.