Pith. sign in

REVIEW 2 major objections 6 minor 1 cited by

A Multi-Centric Anthropomorphic 3D CT Phantom-Based Benchmark Dataset for Harmonization

T0 review · 2 major / 6 minor · reviewed 2026-08-06 · deepseek-v4-flash

Pith's one-line read This paper builds a public benchmark of 1,378 CT image series of the same 3D-printed phantom scanned on 13 clinical CT scanners, so that AI harmonization methods can be developed and compared where patient variation is held fixed.

desk verdict A solid, carefully documented public CT phantom dataset for harmonization benchmarking, with a real caveat: metric sensitivity to harmonization is unverified and one figure caption is contaminated. read the letter →

arxiv 2507.01539 v1 pith:S4K3N62Z submitted 2025-07-02 cs.CV

classification cs.CV
keywords CTharmonizationanthropomorphicphantombenchmarkdatasetdomainshiftradiomicsfeaturestabilitytissueclassificationmulti-centreimage
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

This paper's claim is that scanner-induced domain shift in CT can be turned into a controlled, measurable benchmark by scanning one fixed object many times. The object is a 3D-printed anthropomorphic phantom made of iodine-inked paper that mimics human liver tissue, and it was scanned on 13 clinical CT scanners from four manufacturers at five radiation-dose levels, producing 649 CT scans and 1,378 reconstructed image series in which anatomy never varies. Because the phantom is identical in every acquisition, all remaining differences between image series must come from the scanner and reconstruction pipeline — exactly the variation that AI harmonization aims to remove. The paper contributes the public dataset, an evaluation protocol (image-similarity metrics, feature stability via the intraclass correlation coefficient, and four-class liver-tissue classification with leave-one-scanner-out cross-validation), and baseline results computed without any harmonization, to serve as reference numbers. If the dataset works as intended, it gives the many competing harmonization techniques a common yardstick for what counts as improvement.

What carries the argument

The central object is the anthropomorphic phantom: a 3D print of a real human CT scan in which iodine ink injected into paper raises its attenuation to match liver tissue, accompanied by a thoracic segment and synthetic test patterns, and carrying six annotated liver ROIs from four tissue classes (two cysts, a hemangioma, a metastasis, and two normal regions). Around the phantom sits the harmonized acquisition protocol — acquisition and reconstruction parameters averaged from a survey of clinical thoracoabdominal CT protocols — which fixes tube voltage, pitch, rotation time, collimation, dose, and reconstruction settings as closely as vendor limits allow, so that the scanner becomes the main free variable. The evaluation machinery has three parts: image-level similarity via root mean square error, peak signal-to-noise ratio, and structural similarity between rigidly registered scans; feature-level stability via the intraclass correlation coefficient (ICC(3,1)), which measures feature variation across scanners relative to variation within a scanner; and four-class liver tissue classification on fixed image patches under two protocols — leave-one-scanner-out cross-validation, which tests generalization to scanners never seen in training, and 10-fold cross-validation, where all scanners appear in the training set.

What would settle it

Take a harmonization method that removes per-manufacturer clustering on the phantom and apply it to a small cohort of real patients scanned on two of the same scanners with the same protocol: if the method succeeds on the phantom but leaves patient feature clusters intact, the phantom does not capture the scanner effects the benchmark claims to model. A cheaper check settles the coverage question by inspection: measure the phantom's attenuation range and compare it with the soft-tissue and low-density values, such as lung, that clinical harmonization must handle.

Watch

Extended reading notes

Core claim

The paper establishes a test-retest benchmark: the same physical phantom, scanned repeatedly under controlled settings, so that scanner-related variation is separated from all patient-related variation. The authors show that the benchmark captures real domain shift by demonstrating that features from the phantom's six liver ROIs — handcrafted radiomics, a shallow CNN trained on organ recognition, and a transformer pre-trained on 3D CT volumes — cluster by scanner manufacturer in low-dimensional projections, with reconstruction technique adding a second source of spread. Their baselines indicate the shift exists but does not yet break the provided task: liver-tissue classification accuracy is high even when the test scanner was absent from training, and global structural similarity across scanners is high. The paper reads these observations as showing that the four-class classification task is too easy to expose the benefits of harmonization, and that harmonization effects should be evaluated on target tissues rather than on whole volumes. The contribution is the dataset, the evaluation methodology, and the reference numbers — not a new harmonization algorithm.

Load-bearing premise

The load-bearing premise, stated by the authors in the Methods section, is that the iodine-ink paper phantom reproduces the scanner effects that matter in real patients: its attenuation cannot go below that of paper, so low-density structures such as lung are absent, and it has no anatomical or pathological variability, so harmonization methods tuned on the phantom could still fail on patient data.

Editorial extensions

If this is right

  • Harmonization methods can be benchmarked against the published baselines: improvement means better image-level similarity, higher feature ICC, and higher leave-one-scanner-out classification accuracy.
  • Because anatomy, physiology, and disease are fixed across all acquisitions, a reduction in cross-scanner differences achieved on the phantom is attributable to genuine scanner effects rather than to patient variation.
  • The five dose levels and three reconstruction families (filtered backprojection, iterative reconstruction, and deep-learning reconstruction) allow harmonization across dose and reconstruction to be assessed alongside manufacturer effects.
  • The near-saturated baseline classification implies that the four-class liver task alone will not separate good from bad harmonization methods; the authors suggest adding further tissue classes from the phantom, such as organs or bone.
  • The authors caution that the dataset is not recommended for developing segmentation models, despite providing masks, because of the phantom's missing anatomical diversity and the task's simplicity.

Reading between the lines

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

  • The most informative outcome for this benchmark would be a negative one: if a harmonization method that removes manufacturer clustering on the phantom fails to do so on real multi-centre patient data, the failure would pinpoint which scanner effects the phantom cannot reproduce, most likely in low-density structures.
  • A natural extension is to convert the phantom's thoracic segment and synthetic test patterns into additional classification targets, since the four-class liver task is nearly saturated and cannot distinguish between competing harmonization methods.
  • Because the authors find high global SSIM but strong per-manufacturer feature clustering, harmonization evaluation on this dataset is best restricted to the liver ROIs, where the scanner effects actually appear, rather than to whole volumes.
  • The phantom's paper-density floor limits its attenuation range, so the benchmark says nothing about scanner effects on lung and other low-density tissue; a direct probe would be to insert a low-density calibration insert of known attenuation and re-scan a subset of the 13 scanners.
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

2 major / 6 minor

Summary. Amirian et al. present a publicly available benchmark dataset for evaluating AI harmonization methods in CT imaging. The dataset consists of 1,378 CT image series acquired from a single 3D-printed iodine-ink anthropomorphic phantom on 13 scanners from four manufacturers at eight Swiss institutions, using a harmonized acquisition protocol and five dose levels. The paper describes the phantom, the acquisition protocol, scanner-specific parameters, the data organization, and an open-source code repository. Baseline evaluations without harmonization are reported at three levels: image-level similarity (RMSE, PSNR, SSIM), feature-level stability (ICC), and accuracy of a four-class liver tissue classification task using radiomics and two deep-learning feature extractors under 10-fold and leave-one-scanner-out cross-validation.

Significance. If the benchmark is validated, it will be a valuable community resource: the use of a fixed physical phantom eliminates anatomical and physiological confounds, allowing controlled study of scanner- and dose-induced distribution shifts; the dataset is large (1,378 series), multi-vendor, and public on TCIA; the acquisition protocol is described in detail; and the series counts are internally consistent (Table 2 sums to 1,378). The open-source code and predefined splits lower the barrier for future comparisons. The main weakness is that the evaluation protocol's sensitivity to harmonization is not yet demonstrated.

major comments (2)
  1. [Technical Validation, Table 7 and surrounding text] The central claim of the paper is that the dataset can be used as a benchmark for developing and comparing AI harmonization methods. A benchmark requires that the proposed evaluation metrics have dynamic range to rank methods. This is not demonstrated. In Table 7, the liver tissue classification task is essentially saturated: 10-fold CV accuracy is 0.997±0.001 for radiomics, 1.000±0.000 for shallow CNN, and 0.998±0.002 for SwinUNETR, and LOSO with 12 training scanners reaches 0.985–0.998. A harmonization method cannot improve accuracy beyond 1.0, so this metric cannot discriminate between harmonization methods. The authors acknowledge this ('all the models performed almost perfectly') and propose to add tissue classes, but the released benchmark does not include such a task. To support the central claim, the paper should either add a more challenging task to the released benchmark or include at least one harmonization baseline (e.g., histogram-based image harmonization or ComBat) that demonstrates that the proposed metrics respond in the expected direction.
  2. [Technical Validation, final paragraph and Tables 3–5] The image-level metrics RMSE, PSNR, and SSIM are proposed as benchmark measures, yet the paper provides no evidence that they are sensitive to harmonization. Table 5 reports SSIM values mostly above 0.95, and the text states that 'global similarity measures do not seem to well capture inter-scanner differences.' Since no harmonization method is evaluated, it is unknown whether these metrics change monotonically or meaningfully when a harmonization method is applied. Please compute the metrics on the liver region of interest (rather than the whole phantom volume) and/or demonstrate with a simple harmonization step that the metrics improve.
minor comments (6)
  1. [After Figure 1] The manuscript contains an unrelated passage beginning '1008 D. Groheux et al. Figure 4. Pre-surgery imaging...' inserted between Figure 1 and Figure 2; this appears to be text from a different publication and must be removed.
  2. [Methods, Eq. (2)] The PSNR definition uses max(Ir, Is), the maximum of the two specific image series, whereas PSNR is normally defined with the system's dynamic range; the later statement that L=3000 HU should be used in the PSNR definition for consistency.
  3. [Figure 5 caption] The Figure 5 caption says the UMAP embedding was 'optimized over 100 epochs' while the main text and Figure 6 caption say 1000 epochs; please correct the inconsistency.
  4. [Abstract] The phrase 'which purpose is' should be 'whose purpose is'.
  5. [Methods, Anthropomorphic Phantom and Usage Notes] The abstract and usage notes should state explicitly that the benchmark is scoped to liver/soft-tissue harmonization, since the phantom's minimum attenuation is that of paper and does not cover air-containing structures such as lung; the Methods section does mention this, but the broader wording in the abstract may mislead readers.
  6. [Technical Validation, Eq. (4)] Please clarify the unit of analysis for the ICC: are the six ROIs used as targets and the 13 scanners as raters, and are ICCs averaged across ROIs or across features? This affects the interpretation of the standard deviations in Table 7.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity found: the paper reports empirical baseline measurements on a physical phantom dataset, with no derivation that reduces to its own inputs.

full rationale

This is a dataset and benchmark paper rather than a derivation or prediction paper. The central deliverables are the CT acquisition corpus and baseline measurements: image-level similarity (RMSE, PSNR, SSIM), feature-level stability (ICC), and liver tissue classification accuracy. All of these are computed directly from the acquired scans and expert-defined ROI masks using standard, externally defined formulas (Eqs. 1-4), not from fitted parameters that are then relabeled as predictions. No quantity that the paper claims to establish is defined in terms of the quantity itself, and no result is forced by an ansatz or by a self-citation chain. The self-citations to Bach et al. [6] for the phantom and to Jimenez-del-Toro et al. [17] for the masks are provenance for physical materials and annotations, not load-bearing mathematical premises. The acknowledged phantom limitation (minimum attenuation of paper, inability to represent lung) is a generalizability concern, not a circularity. Likewise, the near-ceiling classification accuracy in Table 7 is a benchmark-sensitivity concern that the paper itself openly discusses ('the classification task is limited in terms of complexity'), but it does not create a logical loop between the paper's claims and its evidence. The paper is self-contained with respect to external validation: the raw data is released on TCIA and the baseline code is open source, so the reported measurements are independently checkable against the physical phantom acquisitions. No circular step can be quoted, and therefore the appropriate score is 0.

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

The paper introduces no new theoretical entities. Its central resource is a physical phantom and a dataset; the only hand-chosen quantities are baseline model hyperparameters and patch sampling choices, which do not affect the dataset's validity. The main assumptions are domain-level: that the phantom fixes anatomy, that expert masks are correct, and that registration and resampling do not distort the measured scanner effects.

free parameters (4)
  • MLP hidden layer sizes = 100, 60, 30
    Hand-chosen architecture for the baseline classifier; not fitted to data and does not affect dataset construction.
  • MLP dropout rate = 0.2
    Hand-chosen for the baseline classifier; used across all classification experiments.
  • Patch size for feature extraction = 32x32x16 voxels
    Hand-chosen patch dimensions for radiomics and DL feature extraction; influences feature values and classification results.
  • Number of patch centers per ROI = 10
    Randomly selected voxel centers per ROI; kept fixed across image series; affects sampling of features.
assumptions (5)
  • domain assumption The phantom is a static, unchanging object so that all image differences reflect scanner and protocol effects.
    Methods section: 'any diversity in reconstructed images, and computed results only reflect the impact of the scanners or acquisition protocol on the images'.
  • domain assumption The expert-defined ROI masks and four tissue classes on the phantom are correct ground truth.
    Methods section: masks from human experts [17], used to define the classification task.
  • domain assumption Rigid registration does not materially bias the computed image similarity metrics.
    Technical Validation acknowledges registration 'has an impact on image appearance' but still uses it before computing RMSE, PSNR, and SSIM.
  • domain assumption The survey-based averaged acquisition parameters represent typical clinical thoracoabdominal CT protocols.
    Methods section: 'a harmonized protocol based on averaged parameters was derived' from a survey of 21 scanners.
  • domain assumption CT intensity values are calibrated and comparable across scanners after resampling to a common voxel spacing and alignment.
    Technical Validation resamples all volumes to 0.6836x0.6836 mm and 2 mm slice thickness before feature extraction.

how reviews work

0 comments
Cite this review

Pith. "Pith review of A Multi-Centric Anthropomorphic 3D CT Phantom-Based Benchmark Dataset for Harmonization." pith.science (2026). https://pith.science/paper/S4K3N62Z

@misc{pith2026250701539,
  author       = {Pith},
  title        = {Pith review of: A Multi-Centric Anthropomorphic 3D CT Phantom-Based Benchmark Dataset for Harmonization},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/S4K3N62Z}},
  note         = {Machine review of arXiv:2507.01539}
}
read the original abstract

Artificial intelligence (AI) has introduced numerous opportunities for human assistance and task automation in medicine. However, it suffers from poor generalization in the presence of shifts in the data distribution. In the context of AI-based computed tomography (CT) analysis, significant data distribution shifts can be caused by changes in scanner manufacturer, reconstruction technique or dose. AI harmonization techniques can address this problem by reducing distribution shifts caused by various acquisition settings. This paper presents an open-source benchmark dataset containing CT scans of an anthropomorphic phantom acquired with various scanners and settings, which purpose is to foster the development of AI harmonization techniques. Using a phantom allows fixing variations attributed to inter- and intra-patient variations. The dataset includes 1378 image series acquired with 13 scanners from 4 manufacturers across 8 institutions using a harmonized protocol as well as several acquisition doses. Additionally, we present a methodology, baseline results and open-source code to assess image- and feature-level stability and liver tissue classification, promoting the development of AI harmonization strategies.

Figures

Figures reproduced from arXiv: 2507.01539 by the authors.

Figure 1
Figure 1. Visual representation of the goal of harmonization techniques as de [PITH_FULL_IMAGE:figures/full_fig_p018_1.png] view at source ↗
Figure 2
Figure 2. Axial and sagittal views of the three segments of the anthropomorphic [PITH_FULL_IMAGE:figures/full_fig_p019_2.png] view at source ↗
Figure 3
Figure 3. Visual overview of the anthropomorphic CT phantom used in this [PITH_FULL_IMAGE:figures/full_fig_p020_3.png] view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Visual comparison of the texture in registered CT series reconstructed [PITH_FULL_IMAGE:figures/full_fig_p021_4.png]
Figure 5
Figure 5. Figure 5: UMAP representation of radiomics features and features derived from [PITH_FULL_IMAGE:figures/full_fig_p022_5.png]
Figure 6
Figure 6. Figure 6: UMAP representation of radiomics features and features derived from [PITH_FULL_IMAGE:figures/full_fig_p023_6.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Distribution Steering via Sliced Optimal Transport Control

    math.OC 2026-08 conditional novelty 6.0 of 10

    A finite-horizon feedback law built from projected one-dimensional optimal transport maps steers distributions, with Gaussian terminal convergence, randomized-to-average convergence, and exact finite-step realization ...

Reference graph

Works this paper leans on

39 extracted references · 36 canonical work pages · cited by 1 Pith paper

  1. [1]

    Decoding tumour phenotype by noninvasive imaging using a quantitative radiomics approach.Nat Com- mun, 5:4006, 6 2014

    Hugo J W L Aerts, Emmanuel Rios Velazquez, Ralph T H Leijenaar, Chintan Parmar, Patrick Grossmann, Sara Carvalho, Johan Bussink, René Monshouwer, Benjamin Haibe-Kains, Derek Rietveld, Frank Hoebers, Michelle M Rietbergen, C René Leemans, Andre Dekker, John Quacken- bush, Robert J Gillies, and Philippe Lambin. Decoding tumour phenotype by noninvasive imagi...

  2. [2]

    Prepnet: A convolutional auto-encoder to homogenize CT scans for cross-dataset medical image analysis

    Mohammadreza Amirian, Javier A Montoya-Zegarra, Jonathan Gruss, Yves D Stebler, Ahmet Selman Bozkir, Marco Calandri, Friedhelm Schwenker, and Thilo Stadelmann. Prepnet: A convolutional auto-encoder to homogenize CT scans for cross-dataset medical image analysis. In2021 14th International Congress on Image and Signal Processing, BioMedical Engineering and ...

  3. [3]

    Learning cross-protocol radiomics and deep feature standardization from CT im- ages of texture phantoms

    Vincent Andrearczyk, Adrien Depeursinge, and Henning Müller. Learning cross-protocol radiomics and deep feature standardization from CT im- ages of texture phantoms. InMedical Imaging 2019: Imaging Informatics for Healthcare, Research, and Applications, volume 10954, pages 109–116. SPIE, 2019

  4. [4]

    Neu- ral network training for cross-protocol radiomic feature standardization in computed tomography

    Vincent Andrearczyk, Adrien Depeursinge, and Henning Müller. Neu- ral network training for cross-protocol radiomic feature standardization in computed tomography. Journal of Medical Imaging, 6(2):024008–024008, 2019

  5. [5]

    Pilot study for the assessment of the best radiomic features for bosniak cyst classification using phantom and radiologist inter- observer selection

    María Aymerich, Mercedes Riveira-Martín, Alejandra García-Baizán, Mar- iña González-Pena, Carmen Sebastià, Antonio López-Medina, Alicia Mesa- Álvarez, Gonzalo Tardágila de la Fuente, Marta Méndez-Castrillón, Andrea Berbel-Rodríguez, et al. Pilot study for the assessment of the best radiomic features for bosniak cyst classification using phantom and radiol...

  6. [6]

    3D-printed iodine-ink CT phan- tom for radiomics feature extraction-advantages and challenges.Medical Physics, 50(9):5682–5697, 2023

    Michael Bach, Christoph Aberle, Adrien Depeursinge, Oscar Jimenez-del Toro, Roger Schaer, Kyriakos Flouris, Ender Konukoglu, Henning Müller, Bram Stieltjes, and Markus M Obmann. 3D-printed iodine-ink CT phan- tom for radiomics feature extraction-advantages and challenges.Medical Physics, 50(9):5682–5697, 2023

  7. [7]

    On various intraclass correlation reliability coefficients

    John J Bartko. On various intraclass correlation reliability coefficients. Psychological bulletin, 83(5):762, 1976. 13

  8. [8]

    Medical Image Harmonization Using Deep Learning Based Canonical Mapping: Toward Robust and Generalizable Learning in Imaging

    Vishnu Bashyam, Jimit Doshi, Guray Erus, Dhivya Srinivasan, Ahmed Abdulkadir, Mohamad Habes, Yong Fan, Colin L. Masters, Paul Maruff, Chuanjun Zhuo, Henry Völzke, Sterling C. Johnson, Jurgen Fripp, Nikolaos Koutsouleris, Theodore Daniel Satterthwaite, Daniel H. Wolf, Raquel E. Gur, Ruben C. Gur, John C. Morris, Marilyn S. Albert, Hans Jörgen Grabe, Susan ...

Show all 39 references
  1. [9]

    An image is worth 16x16 words: Transformers for image recognition at scale

    Alexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn, Xiaohua Zhai, Thomas Unterthiner, Mostafa Dehghani, Matthias Min- derer, Georg Heigold, Sylvain Gelly, Jakob Uszkoreit, and Neil Houlsby. An image is worth 16x16 words: Transformers for image recognition ...

  2. [10]

    An annotated test-retest collection of prostate multiparametric MRI.Scientific Data 2018 5:1, 5:1–13, 12 2018

    Andriy Fedorov, Michael Schwier, David Clunie, Christian Herz, Steve Pieper, Ron Kikinis, Clare Tempany, and Fiona Fennessy. An annotated test-retest collection of prostate multiparametric MRI.Scientific Data 2018 5:1, 5:1–13, 12 2018

  3. [11]

    Radiomics: images are more than pictures, they are data.Radiology, 278(2):563–577, 2016

    Robert J Gillies, Paul E Kinahan, and Hedvig Hricak. Radiomics: images are more than pictures, they are data.Radiology, 278(2):563–577, 2016

  4. [12]

    Eval- uation of domain generalization and adaptation on improving model ro- bustness to temporal dataset shift in clinical medicine.Scientific reports, 12(1):2726, 2022

    Lin Lawrence Guo, Stephen R Pfohl, Jason Fries, Alistair EW Johnson, Jose Posada, Catherine Aftandilian, Nigam Shah, and Lillian Sung. Eval- uation of domain generalization and adaptation on improving model ro- bustness to temporal dataset shift in clinical medicine.Scientific...

  5. [13]

    The many faces of robustness: A critical analysis of out-of- distribution generalization

    Dan Hendrycks, Steven Basart, Norman Mu, Saurav Kadavath, Frank Wang, Evan Dorundo, Rahul Desai, Tyler Zhu, Samyak Parajuli, Mike Guo, et al. The many faces of robustness: A critical analysis of out-of- distribution generalization. InProceedings of the IEEE/CVF international c...

  6. [14]

    Gaussian error linear units (gelus)

    Dan Hendrycks and Kevin Gimpel. Gaussian error linear units (gelus). arXiv preprint arXiv:1606.08415, 2016

  7. [15]

    Phantom-based radiomics feature test–retest stability analysis on photon-counting detector CT.Eu- ropean Radiology, 33(7):4905–4914, 2023

    AlexanderHertel, HishanTharmaseelan, LukasTRotkopf, DominikNören- berg, Philipp Riffel, Konstantin Nikolaou, Jakob Weiss, Fabian Bamberg, Stefan O Schoenberg, Matthias F Froelich, et al. Phantom-based radiomics feature test–retest stability analysis on photon-counting detector...

  8. [16]

    Radiopaque three- dimensional printing: a method to create realistic CT phantoms.Radiology, 282(2):569–575, 2017

    Paul Jahnke, Felix RP Limberg, Andreas Gerbl, Gracia L Ardila Pardo, Victor PB Braun, Bernd Hamm, and Michael Scheel. Radiopaque three- dimensional printing: a method to create realistic CT phantoms.Radiology, 282(2):569–575, 2017. 14

  9. [17]

    Oscar Jimenez-del Toro, Christoph Aberle, Michael Bach, Roger Schaer, Markus M Obmann, Kyriakos Flouris, Ender Konukoglu, Bram Stieltjes, Henning Müller, and Adrien Depeursinge. The discriminative power and stability of radiomics features with computed tomography variations: t...

  10. [18]

    Obmann, André Anjos, Henning Müller, and Adrien Depeursinge

    Oscar Jimenez-del Toro, Christoph Aberle, Roger Schaer, Michael Bach, Kyriakos Flouris, Ender Konukoglu, Bram Stieltjes, Markus M. Obmann, André Anjos, Henning Müller, and Adrien Depeursinge. Comparing sta- bility and discriminatory power of hand-crafted versus deep radiomics:...

  11. [19]

    Radiomics: the bridge between medical imaging and personalized medicine

    Philippe Lambin, Ralph TH Leijenaar, Timo M Deist, Jurgen Peer- lings, Evelyn EC De Jong, Janita Van Timmeren, Sebastian Sanduleanu, Ruben THM Larue, Aniek JG Even, Arthur Jochems, et al. Radiomics: the bridge between medical imaging and personalized medicine. Nature reviews C...

  12. [20]

    Gradient- based learning applied to document recognition.Proceedings of the IEEE, 86(11):2278–2324, 1998

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

  13. [21]

    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. InMedical Image Computing and Computer Assisted Intervention–MICCAI 2021: 24th In- t...

  14. [22]

    Learning disentangled representations in the imag- ing domain

    Xiao Liu, Pedro Sanchez, Spyridon Thermos, Alison Q O’Neil, and Sotirios A Tsaftaris. Learning disentangled representations in the imag- ing domain. Medical Image Analysis, 80:102516, 2022

  15. [23]

    Measuring computed tomography scanner variability of ra- diomics features

    Dennis Mackin, Xenia Fave, Lifei Zhang, David Fried, Jinzhong Yang, Brian Taylor, Edgardo Rodriguez-Rivera, Cristina Dodge, Aaron Kyle Jones, et al. Measuring computed tomography scanner variability of ra- diomics features. Investigative radiology, 50(11):757–765, 2015

  16. [24]

    Making radiomics more reproducible across scanner and imaging protocol variations: a review of harmonization methods

    Shruti Atul Mali, Abdalla Ibrahim, Henry C Woodruff, Vincent Andrea- rczyk, Henning Müller, Sergey Primakov, Zohaib Salahuddin, Avishek Chatterjee, and Philippe Lambin. Making radiomics more reproducible across scanner and imaging protocol variations: a review of harmonization...

  17. [25]

    Generative adversarial networks 15 improve the reproducibility and discriminative power of radiomic features

    Sandra Marcadent, Jeremy Hofmeister, Maria Giulia Preti, Steve P Martin, Dimitri Van De Ville, and Xavier Montet. Generative adversarial networks 15 improve the reproducibility and discriminative power of radiomic features. Radiology: Artificial Intelligence, 2(3):e190035, 2020

  18. [26]

    Deep learning reconstruction improves radiomics feature stability and discriminative power in abdominal CT imaging: a phantom study

    Florian Michallek, Ulrich Genske, Stefan Markus Niehues, Bernd Hamm, and Paul Jahnke. Deep learning reconstruction improves radiomics feature stability and discriminative power in abdominal CT imaging: a phantom study. European Radiology, 32:4587–4595, 7 2022

  19. [27]

    Exploring generalization in deep learning

    Behnam Neyshabur, Srinadh Bhojanapalli, David McAllester, and Nati Srebro. Exploring generalization in deep learning. Advances in neural information processing systems, 30, 2017

  20. [28]

    A guide to combat harmonization of imaging biomarkers in multi- center studies

    Fanny Orlhac, Jakoba J Eertink, Anne-Ségolène Cottereau, Josée M Zijl- stra, Catherine Thieblemont, Michel Meignan, Ronald Boellaard, and Irène Buvat. A guide to combat harmonization of imaging biomarkers in multi- center studies. Journal of Nuclear Medicine, 63(2):172–179, 2022

  21. [29]

    Radiomics: the facts and the challenges of image analysis.European radiology experi- mental, 2:1–8, 2018

    Stefania Rizzo, Francesca Botta, Sara Raimondi, Daniela Origgi, Cristiana Fanciullo, Alessio Giuseppe Morganti, and Massimo Bellomi. Radiomics: the facts and the challenges of image analysis.European radiology experi- mental, 2:1–8, 2018

  22. [30]

    Intraclass correlations: uses in assessing rater reliability.Psychological bulletin, 86(2):420, 1979

    Patrick E Shrout and Joseph L Fleiss. Intraclass correlations: uses in assessing rater reliability.Psychological bulletin, 86(2):420, 1979

  23. [31]

    Alzheimer’s disease classification accuracy is improved by MRI harmonization based on attention-guided generative adversarial networks

    Surabhi Sinha, Sophia I Thomopoulos, Pradeep Lam, Alexandra Muir, and Paul M Thompson. Alzheimer’s disease classification accuracy is improved by MRI harmonization based on attention-guided generative adversarial networks. In 17th international symposium on medical information...

  24. [32]

    Roth, Bennett Land- man, Daguang Xu, Vishwesh Nath, and Ali Hatamizadeh

    Yucheng Tang, Dong Yang, Wenqi Li, Holger R. Roth, Bennett Land- man, Daguang Xu, Vishwesh Nath, and Ali Hatamizadeh. Self-supervised pre-training of swin transformers for 3D medical image analysis. In 2022 IEEE/CVF Conference on Computer Vision and Pattern Recogni- tion (CVPR...

  25. [33]

    Unsupervised MRI homogenization: ap- plication to pediatric anterior visual pathway segmentation

    Carlos Tor-Diez, Antonio Reyes Porras, Roger J Packer, Robert A Avery, and Marius George Linguraru. Unsupervised MRI homogenization: ap- plication to pediatric anterior visual pathway segmentation. InMachine Learning in Medical Imaging: 11th International Workshop, MLMI 2020, ...

  26. [34]

    Visualizing non-metric sim- ilarities in multiple maps.Machine Learning, 87(1):33–55, 2012

    Laurens Van der Maaten and Geoffrey Hinton. Visualizing non-metric sim- ilarities in multiple maps.Machine Learning, 87(1):33–55, 2012. 16

  27. [35]

    Compu- tational radiomics system to decode the radiographic phenotype.Cancer Research, 77:e104–e107, 2017

    Joost J M van Griethuysen, Andriy Fedorov, Chintan Parmar, Ahmed Hosny, Nicole Aucoin, Vivek Narayan, Regina G H Beets-Tan, Jean- Christophe Fillion-Robin, Steve Pieper, and Hugo J W L Aerts. Compu- tational radiomics system to decode the radiographic phenotype.Cancer Research...

  28. [36]

    Radiomic feature robustness evaluations in ultrasound imaging

    Jing Wang, Boran Zhou, Xiaofeng Yang, and Tian Liu. Radiomic feature robustness evaluations in ultrasound imaging. In Medical Imaging 2023: Ultrasonic Imaging and Tomography, volume 12470, pages 159–164. SPIE, 2023

  29. [37]

    Contrastive cross-site learning with redesigned net for COVID-19 CT classification.IEEE Journal of Biomed- ical and Health Informatics, 24(10):2806–2813, 2020

    Zhao Wang, Quande Liu, and Qi Dou. Contrastive cross-site learning with redesigned net for COVID-19 CT classification.IEEE Journal of Biomed- ical and Health Informatics, 24(10):2806–2813, 2020

  30. [38]

    Image quality assessment: from error visibility to structural similarity

    Zhou Wang, Alan C Bovik, Hamid R Sheikh, and Eero P Simoncelli. Image quality assessment: from error visibility to structural similarity. IEEE Transactions on Image Processing, 13(4):600–612, 2004

  31. [39]

    Radiomic feature repeatability and its impact on prognostic model generalizability: A multi-institutional study on nasopha- ryngeal carcinoma patients

    Jiang Zhang, Sai Kit Lam, Xinzhi Teng, Zongrui Ma, Xinyang Han, Yuan- peng Zhang, Andy Lai Yin Cheung, Tin Ching Chau, Sherry Chor Yi Ng, Francis Kar Ho Lee, Kwok Hung Au, Celia Wai Yi Yip, Victor Ho Fun Lee, Ying Han, and Jing Cai. Radiomic feature repeatability and its impac...

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.