Pith. sign in

REVIEW 3 major objections 3 minor 82 references

This paper claims that particle instance segmentation can be trained entirely without human annotations by using cross-scan consistency from reshuffled scans as a correctness signal.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

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

A self-training method uses cross-scan particle matching instead of human labels to train and evaluate 3D particle instance segmentation, reporting 97 percent volume coverage and 54,000+ particles on quartz fragments.

T0 review reviewed 2026-08-05 challenge →

load-bearing objection Clever consistency-based self-training that lacks an external correctness anchor; the claimed accuracy is circular until validated against ground truth. the 3 major comments →

arxiv 2508.16224 v1 pith:HVXNEW4N submitted 2025-08-22 eess.IV cs.CV

Self-Validated Learning for Particle Separation: A Correctness-Based Self-Training Framework Without Human Labels

Philipp D. L\"osel , Aleese Barron , Yulai Zhang , Matthias Fabian , Benjamin Young , Nicolas Francois , Andrew M. Kingston This is my paper
classification eess.IV cs.CV
keywords self-validated learningself-traininginstance segmentationparticle separationX-ray micro-CTpseudo-labelscross-scan consistencylabel-free evaluation
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

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 attempts to remove the last human step in training instance-segmentation models for particulate samples: label creation. It proposes self-validated learning, in which the same physical particles are scanned several times after being reshuffled, and the instances that appear consistently across all scans are treated as correct labels for retraining. A correctness check based on cross-scan instance matching replaces confidence thresholds, and implicit boundary detection makes it possible to learn from partially labeled volumes. On crushed quartz, the loop reaches over 97% volume coverage and 54,000 identified particles after three iterations, entirely without manual annotations. If the premise holds, the same mechanism also gives a ground-truth-free way to evaluate and compare segmentation models.

Core claim

The central claim is that a segmentation model can learn to separate touching particles in 3D tomographic scans without any human annotation, provided the same physical sample can be reshuffled and rescanned multiple times. The method registers and matches particle instances across these independently reshuffled volumes; instances found consistently in all scans are treated as self-validated pseudo-labels and added to the training set, while unmatched or inconsistent regions are ignored. Implicit boundary detection lets the network train from this partially labeled data by supervising only confidently segmented inner regions. After three iterations on crushed quartz sphere fragments, the app

What carries the argument

Cross-scan instance matching with implicit boundary detection. The self-validation score is the mechanism: after each reshuffled rescan, candidate instances are rotationally aligned and matched; a particle that matches across all scans is considered correct and becomes a pseudo-label. Implicit boundary detection restricts supervision to confidently segmented regions, so the partially labeled volumes can be used for training without needing full masks.

Load-bearing premise

The load-bearing premise is that a particle appearing consistently across reshuffled scans is segmented correctly; if the same error repeats in every scan—for example, two touching particles always merged—the matching criterion will certify a wrong label.

What would settle it

A synthetic phantom with known ground truth: reshuffle and rescan it, run the loop, and compare every accepted pseudo-label with the known truth. If a deliberately induced, reshuffle-invariant error (such as a consistent merge of two touching particles) still receives high consistency, the self-validation is not measuring correctness.

Watch this falsifier. Get emailed when new claim-graph text bears on it.

If this is right

  • Training sets for particle instance segmentation can be built from unlabeled scans alone, removing the annotation bottleneck for new materials.
  • The same consistency score can evaluate models, tune hyperparameters, and monitor imaging quality without ground-truth labels.
  • Self-training with correctness-based filtering should avoid the noise-accumulation collapse that plagues confidence-threshold pseudo-labeling.
  • The demonstration on more than 54,000 particles in quartz scans indicates the loop scales to large, densely packed tomographic volumes.
  • Because the method is model-agnostic, it can be applied to any segmentation network or existing segmenter, not only the one used in the experiments.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The consistency criterion is only as strong as the physical independence of the rescans; if reshuffling preserves the same ambiguous contact geometry, consistently wrong merges will be certified as correct.
  • For samples that cannot be reshuffled—fixed geological cores, in-vivo specimens, radiation-sensitive materials—the cross-scan premise fails, so some other independent observation or perturbation would be needed.
  • A direct stress test would inject synthetic merge errors into one scan and check that the consistency score drops; this could be done with existing synthetic particle pack generators without any manual labels.
  • The same consistency logic could extend beyond particles to any population of discrete objects that can be physically perturbed and re-imaged, and could also grade the confidence of pretrained segmenters before trusting their outputs.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 3 minor

Summary. The paper proposes Self-Validated Learning (SVL), a self-training framework for 3D particle instance segmentation that claims to require no human annotations. Instead of using confidence thresholds, SVL selects pseudo-labels by matching predicted particle instances across multiple independently reshuffled and rescanned volumes of the same sample; the same cross-scan matching is also proposed as an autonomous evaluation metric. The abstract claims that after three iterations the method segments over 97% of total particle volume and identifies more than 54,000 particles in tomographic scans of quartz fragments. The available manuscript text consists of the abstract, introduction, and part of Section II (Related Work, Section II-A), plus references; the methods and experimental sections are absent, so the central quantitative claims cannot be verified from the supplied text.

Significance. If the claims are substantiated, the work would be a valuable contribution to autonomous particle segmentation: it addresses a real bottleneck (manual annotation for micro-CT particle data), proposes an interesting use of physical reshuffling as a training signal, and appears to be integrated into the open-source Biomedisa platform. The idea of consistency across independent scans as a pseudo-label filter is novel and worth developing. However, the significance is entirely conditional: the manuscript as supplied does not contain the method details, the experimental setup, or a non-circular evaluation. The main scientific risk is that the self-validation criterion is a consistency check rather than a correctness check; the paper needs an external anchor to break that circularity before its performance claims can be counted.

major comments (3)
  1. [Abstract and Section I] The central quantitative claims—"over 97% of total particle volume" and "more than 54,000 individual particles" after three iterations—are not backed by any presented experiments, error bars, or sensitivity analysis. The full text supplied ends in the middle of Section II-A, after the Cellpose paragraph, and no methods (Section III) or results (Section IV) are included. The introduction even cross-references the missing "Section IV-E." Without the experimental section, the core claims are unverifiable. This is load-bearing and blocks acceptance.
  2. [Section I, Fig. 1] The selection of pseudo-labels and the autonomous evaluation both rely on the same cross-scan matching signal. Consistency across reshuffled scans is a necessary condition for correct segmentation, not a sufficient one. Systematic errors that reproduce in all reshuffled scans, such as consistently splitting a particle with a deep notch or consistently merging an interlocked pair, will pass the matching test. The claimed 97% volume and 54,000-particle count can be self-consistent while being wrong. The manuscript needs an independent correctness anchor: manually annotated subset, synthetic ground truth, or a metric whose correlation with ground-truth metrics is demonstrated. The phrase "correctness-based selection" overstates what the matching criterion can establish.
  3. [Section I, last paragraph] The claim that the framework "enables fully autonomous model evaluation without the need for ground truth annotations" is not justified. Comparisons with other methods performed under the same consistency metric do not supply an external standard; they only show relative consistency. To support autonomous evaluation, the paper must show that the consistency score tracks a standard segmentation metric (e.g., Dice/ARAND on a labeled subset) and quantify the agreement, including failure cases. Without this, the evaluation is circular.
minor comments (3)
  1. [Introduction, Section IV-E reference] The introduction says "see also our Section IV-E," but Section IV-E is not present in the supplied text. If the full manuscript includes it, ensure the cross-reference is correct; otherwise remove it.
  2. [Abstract and Introduction] The object is described as "crushed quartz sphere fragments" in the Introduction and "tomographic scans of quartz fragments" in the Abstract; please use consistent terminology.
  3. [Abstract] The phrase "over 97% of the total particle volume" lacks a definition of the numerator and denominator. Is this the fraction of ground-truth particle volume covered by matched instances, or the fraction of the model's volume that is matched across scans? Specify the metric.

Circularity Check

1 steps flagged

Cross-scan matching is both the pseudo-label selection criterion and the autonomous evaluation metric, so the reported 97% / 54,000-particle accuracy claim reduces to the self-validation signal by construction.

specific steps
  1. self definitional [Abstract; Section I (Proposed Method)]
    "Our method leverages implicit boundary detection and iteratively refines the training set by identifying particles that can be consistently matched across reshuffled scans of the same sample. This self-validation mechanism mitigates the impact of noisy pseudo-labels, enabling robust learning from unlabeled data. After just three iterations, our approach accurately segments over 97% of the total particle volume and identifies more than 54,000 individual particles in tomographic scans of quartz fragments. Importantly, the framework also enables fully autonomous model evaluation without the need"

    The pseudo-label selector and the final evaluation are the same operation: particles are retained only if 'consistently matched across reshuffled scans,' and that same matching is used to claim 'accurately segments over 97% of the total particle volume' and 'fully autonomous model evaluation.' Since training labels are chosen to maximize cross-scan matchability, the reported performance measures the model's agreement with the matching criterion, not correctness against independent ground truth. Systematic errors stable across reshuffles—consistent splitting of notched particles or merging of interlocked particles—satisfy the criterion and are counted as correct. Section I makes the identification explicit ('correctness-based selection strategy: only predictions that are validated through c

full rationale

The central loop of the paper is: segment each reshuffled scan, rotationally align and match instances, retain only matched instances as pseudo-labels, retrain, and after three iterations report that the model 'accurately segments over 97% of the total particle volume' and 'identifies more than 54,000 particles,' with evaluation performed 'without the need for ground truth annotations.' Both the pseudo-label selection and the final evaluation use the same cross-scan instance matching operation. Therefore the quantity called 'accurate' is, by the paper's own description, the fraction of volume that is consistently matchable across reshuffles—not a quantity anchored to ground truth. Stable systematic errors (e.g., a notched particle consistently split in every reshuffle, or two interlocked particles consistently merged) satisfy the matching criterion and would be counted as correct. The comparisons with state-of-the-art methods are also framed under the same consistency-based evaluation, so they do not provide an independent anchor. This is a genuine circularity in the central claim, though the framework still has independent algorithmic content (iterative self-training, implicit boundary detection), which is why the score is 6 rather than 8-10.

Axiom & Free-Parameter Ledger

1 free parameters · 3 axioms · 0 invented entities

The framework introduces a new training protocol, not new physical entities. The listed assumptions are the ones the central claim depends on; the free parameter count is minimal in the abstract, but the matching threshold and other hyperparameters would need to be audited in the full methods section.

free parameters (1)
  • number of self-training iterations = 3
    The abstract reports convergence after three iterations; this stopping point is chosen by the authors, not derived from first principles.
axioms (3)
  • domain assumption Reshuffled scans of the same sample provide independent observations of the same particle instances.
    Used in the abstract and Fig. 1; the consistency signal is only informative if segmentation errors are conditionally independent across scans, which the paper does not justify.
  • domain assumption Consistency of an instance across reshuffled scans implies the instance is correctly segmented.
    This is the correctness-based selection criterion; it fails when both scans reproduce the same systematic error.
  • domain assumption Training can proceed from partially labeled data using implicit boundary detection, with unlabeled regions excluded from the loss.
    Introduced in the introduction; requires the loss function to ignore unlabeled voxels and assumes the network can learn from partial labels.

reviewed 2026-08-05 · how reviews work

0 comments
Cite this review

Pith. "Pith review of Self-Validated Learning for Particle Separation: A Correctness-Based Self-Training Framework Without Human Labels." pith.science (2026). https://pith.science/paper/HVXNEW4N

@misc{pith2026250816224,
  author       = {Pith},
  title        = {Pith review of: Self-Validated Learning for Particle Separation: A Correctness-Based Self-Training Framework Without Human Labels},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/HVXNEW4N}},
  note         = {Machine review of arXiv:2508.16224}
}
Share X Bluesky LinkedIn Reddit HN
read the original abstract

Non-destructive 3D imaging of large multi-particulate samples is essential for quantifying particle-level properties, such as size, shape, and spatial distribution, across applications in mining, materials science, and geology. However, accurate instance segmentation of particles in tomographic data remains challenging due to high morphological variability and frequent particle contact, which limit the effectiveness of classical methods like watershed algorithms. While supervised deep learning approaches offer improved performance, they rely on extensive annotated datasets that are labor-intensive, error-prone, and difficult to scale. In this work, we propose self-validated learning, a novel self-training framework for particle instance segmentation that eliminates the need for manual annotations. Our method leverages implicit boundary detection and iteratively refines the training set by identifying particles that can be consistently matched across reshuffled scans of the same sample. This self-validation mechanism mitigates the impact of noisy pseudo-labels, enabling robust learning from unlabeled data. After just three iterations, our approach accurately segments over 97% of the total particle volume and identifies more than 54,000 individual particles in tomographic scans of quartz fragments. Importantly, the framework also enables fully autonomous model evaluation without the need for ground truth annotations, as confirmed through comparisons with state-of-the-art instance segmentation techniques. The method is integrated into the Biomedisa image analysis platform (https://github.com/biomedisa/biomedisa/).

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

82 extracted references · 63 canonical work pages · 1 internal anchor

  1. [1]

    A Review of Particle Shape Effects on Material Properties for Various Engineering Applications: From Macro to Nanoscale,

    U. Ulusoy, “A Review of Particle Shape Effects on Material Properties for Various Engineering Applications: From Macro to Nanoscale,” Minerals, vol. 13, no. 1, p. 91, Jan. 2023. [Online]. Available: https://www.mdpi.com/2075-163X/13/1/91

    2023

  2. [2]

    Soil behaviour: The role of particle shape,

    J. C. Santamarina and G. C. Cho, “Soil behaviour: The role of particle shape,” in Advances in geotechnical engineering: The Skempton conference , Jun. 2015, pp. 604–617, eprint: https://www.icevirtuallibrary.com/doi/pdf/10.1680/aigev1.32644.0035. [Online]. Available: https://www.icevirtuallibrary.com/doi/abs/10.1680/ aigev1.32644.0035

    work page arXiv 2015

  3. [3]

    Deep convolutional neural network for 3D mineral identification and liberation analysis,

    K. Tang, Y . D. Wang, P. Mostaghimi, M. Knackstedt, C. Hargrave, and R. T. Armstrong, “Deep convolutional neural network for 3D mineral identification and liberation analysis,” Minerals Engineering , vol. 183, p. 107592, Jun. 2022. [Online]. Available: https://linkinghub.elsevier. com/retrieve/pii/S0892687522002023

    2022

  4. [4]

    Techniques in helical scanning, dynamic imaging and image segmentation for improved quantitative analysis with X-ray micro-CT,

    A. Sheppard, S. Latham, J. Middleton, A. Kingston, G. Myers, T. Varslot, A. Fogden, T. Sawkins, R. Cruikshank, M. Saadatfar, N. Francois, C. Arns, and T. Senden, “Techniques in helical scanning, dynamic imaging and image segmentation for improved quantitative analysis with X-ray micro-CT,” Nuclear Instruments and Methods in Physics Research Section B: Bea...

    work page 2014

  5. [5]

    3D image segmentation for analysis of multisize particles in a packed particle bed,

    Y . Wang, C. Lin, and J. Miller, “3D image segmentation for analysis of multisize particles in a packed particle bed,” Powder Technology, vol. 301, pp. 160–168, Nov. 2016. [Online]. Available: https://linkinghub.elsevier.com/retrieve/pii/S0032591016302339

    work page 2016

  6. [6]

    Improved 3d image segmentation for x-ray tomographic analysis of packed particle beds,

    ——, “Improved 3d image segmentation for x-ray tomographic analysis of packed particle beds,” Minerals Engineering , vol. 83, pp. 185–191, 2015. [Online]. Available: https://www.sciencedirect.com/ science/article/pii/S0892687515300777

    work page 2015

  7. [7]

    Uncertainty-aware particle segmentation for electron microscopy at varied length scales,

    L. Rettenberger, N. J. Szymanski, Y . Zeng, J. Schuetzke, S. Wang, G. Ceder, and M. Reischl, “Uncertainty-aware particle segmentation for electron microscopy at varied length scales,” npj Computational Materials, vol. 10, no. 1, p. 124, Jun. 2024. [Online]. Available: https://www.nature.com/articles/s41524-024-01302-w

    work page 2024

  8. [8]

    A method of blasted rock image segmentation based on improved watershed algorithm,

    Q. Guo, Y . Wang, S. Yang, and Z. Xiang, “A method of blasted rock image segmentation based on improved watershed algorithm,” Scientific Reports, vol. 12, no. 1, p. 7143, May 2022. [Online]. Available: https://www.nature.com/articles/s41598-022-11351-0

    work page 2022

  9. [9]

    Improved particle separation, characterisation and analysis for ore beneficiation studies using 3d x-ray micro-computed tomography,

    Y . Zhang, N. Francois, L. Deakin, A. Barron, C. Evans, S. Bensley, P. D. L ¨osel, N. Kandula, R. Henley, and M. Knackstedt, “Improved particle separation, characterisation and analysis for ore beneficiation studies using 3d x-ray micro-computed tomography,” Minerals Engineering, vol. 216, p. 108835, 2024. [Online]. Available: https: //www.sciencedirect.c...

    work page 2024

  10. [10]

    Particleseg3d: A scalable out-of-the-box deep learning segmentation solution for individual particle characterization from micro ct images in mineral processing and recycling,

    K. Gotkowski, S. Gupta, J. R. Godinho, C. G. Tochtrop, K. H. Maier- Hein, and F. Isensee, “Particleseg3d: A scalable out-of-the-box deep learning segmentation solution for individual particle characterization from micro ct images in mineral processing and recycling,” Powder Technology, vol. 434, p. 119286, 2024. [Online]. Available: https://www.sciencedir...

    work page 2024

  11. [11]

    Cell detection and segmentation in microscopy images with improved mask r-cnn,

    S. Fujita and X.-H. Han, “Cell detection and segmentation in microscopy images with improved mask r-cnn,” in Computer Vision – ACCV 2020 Workshops, I. Sato and B. Han, Eds. Cham: Springer International Publishing, 2021, pp. 58–70

    work page 2020

  12. [12]

    Segment Anything for Microscopy,

    A. Archit, L. Freckmann, S. Nair, N. Khalid, P. Hilt, V . Rajashekar, M. Freitag, C. Teuber, G. Buckley, S. V on Haaren, S. Gupta, A. Dengel, S. Ahmed, and C. Pape, “Segment Anything for Microscopy,” Nature Methods, vol. 22, no. 3, pp. 579–591, Mar. 2025. [Online]. Available: https://www.nature.com/articles/s41592-024-02580-4

    work page 2025

  13. [13]

    Synthetic particle pack generation for augmentation and testing in geological tomographic segmentation,

    B. Wang, A. Kingston, P. D. L ¨osel, and W. Creemers, “Synthetic particle pack generation for augmentation and testing in geological tomographic segmentation,” Tomography of Materials and Structures , p. 100072, 2025. [Online]. Available: https://www.sciencedirect.com/ science/article/pii/S2949673X25000257

    work page 2025

  14. [14]

    Is learning the n-th thing any easier than learning the first?

    S. Thrun, “Is learning the n-th thing any easier than learning the first?” in Proceedings of (NeurIPS) Neural Information Processing Systems , November 1995, pp. 640 – 646

    work page 1995

  15. [15]

    Segment Anything,

    A. Kirillov, E. Mintun, N. Ravi, H. Mao, C. Rolland, L. Gustafson, T. Xiao, S. Whitehead, A. C. Berg, W.-Y . Lo, P. Doll´ar, and R. Girshick, “Segment Anything,” 2023, version Number: 1. [Online]. Available: https://arxiv.org/abs/2304.02643

    Pith/arXiv arXiv 2023

  16. [16]

    Mask R-CNN,

    K. He, G. Gkioxari, P. Doll ´ar, and R. Girshick, “Mask R-CNN,” 2017, version Number: 3. [Online]. Available: https://arxiv.org/abs/1703.06870

    Pith/arXiv arXiv 2017

  17. [17]

    Detectron2,

    Y . Wu, A. Kirillov, F. Massa, W.-Y . Lo, and R. Girshick, “Detectron2,” https://github.com/facebookresearch/detectron2, 2019

    2019

  18. [18]

    Ultralytics yolov8,

    G. Jocher, A. Chaurasia, and J. Qiu, “Ultralytics yolov8,” 2023. [Online]. Available: https://github.com/ultralytics/ultralytics

    2023

  19. [19]

    BioImage Model Zoo: A Community-Driven Resource for Accessible Deep Learning in BioImage Analysis,

    W. Ouyang, F. Beuttenmueller, E. G ´omez-de Mariscal, C. Pape, T. Burke, C. Garcia-L ´opez-de Haro, C. Russell, L. Moya-Sans, C. de-la Torre-Guti´errez, D. Schmidt, D. Kutra, M. Novikov, M. Weigert, U. Schmidt, P. Bankhead, G. Jacquemet, D. Sage, R. Henriques, A. Mu ˜noz-Barrutia, E. Lundberg, F. Jug, and A. Kreshuk, “BioImage Model Zoo: A Community-Drive...

    work page doi:10.1101/2022.06.07 2022

  20. [20]

    DeepImageJ: A user-friendly environment to run deep learning models in ImageJ,

    E. G ´omez-de Mariscal, C. Garc´ıa-L´opez-de Haro, W. Ouyang, L. Donati, E. Lundberg, M. Unser, A. Mu ˜noz-Barrutia, and D. Sage, “DeepImageJ: A user-friendly environment to run deep learning models in ImageJ,” Nature Methods, vol. 18, no. 10, pp. 1192–1195, Oct. 2021. [Online]. Available: https://www.nature.com/articles/s41592-021-01262-9

    work page 2021

  21. [21]

    ilastik: interactive machine learning for (bio)image analysis,

    S. Berg, D. Kutra, T. Kroeger, C. N. Straehle, B. X. Kausler, C. Haubold, M. Schiegg, J. Ales, T. Beier, M. Rudy, K. Eren, J. I. Cervantes, B. Xu, F. Beuttenmueller, A. Wolny, C. Zhang, U. Koethe, F. A. Hamprecht, and A. Kreshuk, “ilastik: interactive machine learning for (bio)image analysis,” Nature Methods, vol. 16, no. 12, pp. 1226–1232, Dec. 2019. [On...

    work page 2019

  22. [22]

    Segment anything in medical images,

    J. Ma, Y . He, F. Li, L. Han, C. You, and B. Wang, “Segment anything in medical images,” Nature Communications , vol. 15, no. 1, p. 654, Jan. 2024. [Online]. Available: https://www.nature.com/articles/ s41467-024-44824-z

    work page 2024

  23. [23]

    A survey on self-supervised learning: Algorithms, applications, and future trends,

    J. Gui, T. Chen, J. Zhang, Q. Cao, Z. Sun, H. Luo, and D. Tao, “A survey on self-supervised learning: Algorithms, applications, and future trends,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 46, no. 12, pp. 9052–9071, 2024

    work page 2024

  24. [24]

    Masked au- toencoders are scalable vision learners,

    K. He, X. Chen, S. Xie, Y . Li, P. Doll ´ar, and R. Girshick, “Masked au- toencoders are scalable vision learners,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , June 2022, pp. 16 000–16 009

    2022

  25. [25]

    Unsupervised representation learning by predicting image rotations,

    S. Gidaris, P. Singh, and N. Komodakis, “Unsupervised representation learning by predicting image rotations,” 2018. [Online]. Available: https://arxiv.org/abs/1803.07728

    Pith/arXiv arXiv 2018

  26. [26]

    Unsupervised learning of visual representations by solving jigsaw puzzles,

    M. Noroozi and P. Favaro, “Unsupervised learning of visual representations by solving jigsaw puzzles,” 2017. [Online]. Available: https://arxiv.org/abs/1603.09246

    Pith/arXiv arXiv 2017

  27. [27]

    Contrastive learning of global and local features for medical image segmentation with limited annotations,

    K. Chaitanya, E. Erdil, N. Karani, and E. Konukoglu, “Contrastive learning of global and local features for medical image segmentation with limited annotations,” 2020. [Online]. Available: https://arxiv.org/ abs/2006.10511

    Pith/arXiv arXiv 2020

  28. [28]

    Pseudo-label : The simple and efficient semi-supervised learning method for deep neural networks,

    D.-H. Lee, “Pseudo-label : The simple and efficient semi-supervised learning method for deep neural networks,” in Workshop on Challenges in Representation Learning , 2013. [Online]. Available: https://api.semanticscholar.org/CorpusID:18507866

    work page 2013

  29. [29]

    Self-training: A survey,

    M.-R. Amini, V . Feofanov, L. Pauletto, L. Hadjadj, ´Emilie Devijver, and Y . Maximov, “Self-training: A survey,” Neurocomputing, vol. 616, p. 128904, 2025. [Online]. Available: https://www.sciencedirect.com/ science/article/pii/S0925231224016758

    work page 2025

  30. [30]

    S4L: Self-Supervised Semi-Supervised Learning,

    X. Zhai, A. Oliver, A. Kolesnikov, and L. Beyer, “S4L: Self-Supervised Semi-Supervised Learning,” in 2019 IEEE/CVF International Conference on Computer Vision (ICCV) . Seoul, Korea (South): IEEE, Oct. 2019, pp. 1476–1485. [Online]. Available: https://ieeexplore.ieee.org/document/9010283/

    work page arXiv 2019

  31. [31]

    Semi-supervised learning (chapelle, o. et al., eds.; 2006) [book reviews],

    O. Chapelle, B. Scholkopf, and A. Zien, Eds., “Semi-supervised learning (chapelle, o. et al., eds.; 2006) [book reviews],” IEEE Transactions on Neural Networks, vol. 20, no. 3, pp. 542–542, 2009

    work page 2006

  32. [32]

    Introducing Biomedisa as an open-source online platform for biomedical image segmentation,

    P. D. L ¨osel, T. van de Kamp, A. Jayme, A. Ershov, T. Farag ´o, O. Pichler, N. Tan Jerome, N. Aadepu, S. Bremer, S. A. Chilingaryan, M. Heethoff, A. Kopmann, J. Odar, S. Schmelzle, M. Zuber, J. Wittbrodt, T. Baumbach, and V . Heuveline, “Introducing Biomedisa as an open-source online platform for biomedical image segmentation,” Nature Communications, vol...

    work page 2020

  33. [33]

    A Computational Approach to Edge Detection,

    J. Canny, “A Computational Approach to Edge Detection,” IEEE Transactions on Pattern Analysis and Machine Intelligence , vol. PAMI-8, no. 6, pp. 679–698, Nov. 1986. [Online]. Available: https://ieeexplore.ieee.org/document/4767851

    work page arXiv 1986

  34. [34]

    Trainable Weka Segmentation: a machine learning tool for microscopy pixel classification,

    I. Arganda-Carreras, V . Kaynig, C. Rueden, K. W. Eliceiri, J. Schindelin, A. Cardona, and H. Sebastian Seung, “Trainable Weka Segmentation: a machine learning tool for microscopy pixel classification,” Bioinfor- matics, vol. 33, no. 15, pp. 2424–2426, Aug. 2017. [Online]. Available: https://academic.oup.com/bioinformatics/article/33/15/2424/3092362 18

    work page 2017

  35. [35]

    Fiji: an open-source platform for biological-image analysis,

    J. Schindelin, I. Arganda-Carreras, E. Frise, V . Kaynig, M. Longair, T. Pietzsch, S. Preibisch, C. Rueden, S. Saalfeld, B. Schmid, J.-Y . Tinevez, D. J. White, V . Hartenstein, K. Eliceiri, P. Tomancak, and A. Cardona, “Fiji: an open-source platform for biological-image analysis,” Nature Methods , vol. 9, no. 7, pp. 676–682, Jul. 2012. [Online]. Availabl...

    work page 2012

  36. [36]

    U-Net: Convolutional Networks for Biomedical Image Segmentation,

    O. Ronneberger, P. Fischer, and T. Brox, “U-Net: Convolutional Networks for Biomedical Image Segmentation,” in Medical Image Com- puting and Computer-Assisted Intervention – MICCAI 2015 , N. Navab, J. Hornegger, W. M. Wells, and A. F. Frangi, Eds. Cham: Springer International Publishing, 2015, vol. 9351, pp. 234–241. [Online]. Available: http://link.sprin...

    doi:10.1007/978-3-319-24574-4 2015

  37. [37]

    Accurate and versatile 3D segmentation of plant tissues at cellular resolution,

    A. Wolny, L. Cerrone, A. Vijayan, R. Tofanelli, A. V . Barro, M. Louveaux, C. Wenzl, S. Strauss, D. Wilson-S ´anchez, R. Lymbouridou, S. S. Steigleder, C. Pape, A. Bailoni, S. Duran- Nebreda, G. W. Bassel, J. U. Lohmann, M. Tsiantis, F. A. Hamprecht, K. Schneitz, A. Maizel, and A. Kreshuk, “Accurate and versatile 3D segmentation of plant tissues at cellul...

    work page 2020

  38. [38]

    Deep Watershed Transform for Instance Segmentation,

    M. Bai and R. Urtasun, “Deep Watershed Transform for Instance Segmentation,” in 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Honolulu, HI: IEEE, Jul. 2017, pp. 2858–

    work page 2017

  39. [39]

    Cellpose: a generalist algorithm for cellular segmentation,

    C. Stringer, T. Wang, M. Michaelos, and M. Pachitariu, “Cellpose: a generalist algorithm for cellular segmentation,” Nature Methods , vol. 18, no. 1, pp. 100–106, Jan. 2021. [Online]. Available: https://www.nature.com/articles/s41592-020-01018-x

    work page 2021

  40. [40]

    Microsoft coco: Common objects in context,

    T.-Y . Lin, M. Maire, S. Belongie, L. Bourdev, R. Girshick, J. Hays, P. Perona, D. Ramanan, C. L. Zitnick, and P. Doll ´ar, “Microsoft coco: Common objects in context,” 2015. [Online]. Available: https://arxiv.org/abs/1405.0312

    Pith/arXiv arXiv 2015

  41. [41]

    Imagenet: A large-scale hierarchical image database,

    J. Deng, W. Dong, R. Socher, L.-J. Li, K. Li, and L. Fei-Fei, “Imagenet: A large-scale hierarchical image database,” in 2009 IEEE Conference on Computer Vision and Pattern Recognition , 2009, pp. 248–255

    2009

  42. [42]

    You Only Look Once: Unified, Real-Time Object Detection,

    J. Redmon, S. Divvala, R. Girshick, and A. Farhadi, “You Only Look Once: Unified, Real-Time Object Detection,” 2015, version Number: 5. [Online]. Available: https://arxiv.org/abs/1506.02640

    Pith/arXiv arXiv 2015

  43. [43]

    A FORMAL PROOF OF THE KEPLER CONJECTURE,

    T. Hales, M. Adams, G. Bauer, T. D. Dang, J. Harrison, L. T. Hoang, C. Kaliszyk, V . Magron, S. Mclaughlin, T. Nguyen, Q. T. Nguyen, T. Nipkow, S. Obua, J. Pleso, J. Rute, A. Solovyev, T. H. A. Ta, N. T. Tran, T. D. Trieu, J. Urban, K. Vu, and R. Zumkeller, “A FORMAL PROOF OF THE KEPLER CONJECTURE,” Forum of Mathematics, Pi, vol. 5, p. e2, 2017. [Online]....

    work page 2017

  44. [44]

    A quantified study of segmentation techniques on synthetic geological XRM and FIB-SEM images,

    M. Andrew, “A quantified study of segmentation techniques on synthetic geological XRM and FIB-SEM images,” Computational Geosciences, vol. 22, no. 6, pp. 1503–1512, Dec. 2018. [Online]. Available: http://link.springer.com/10.1007/s10596-018-9768-y

    work page doi:10.1007/s10596-018-9768-y 2018

  45. [45]

    Particle classification of iron ore sinter green bed mixtures by 3d x- ray microcomputed tomography and machine learning,

    K. Tang, Y . D. Wang, Y . Niu, T. A. Honeyands, D. O. Dea, P. Mostaghimi, R. T. Armstrong, and M. Knackstedt, “Particle classification of iron ore sinter green bed mixtures by 3d x- ray microcomputed tomography and machine learning,” Powder Technology, vol. 415, p. 118151, 2023. [Online]. Available: https: //www.sciencedirect.com/science/article/pii/S0032...

    work page 2023

  46. [46]

    Ore image segmentation method using U-Net and Res unet convolutional networks,

    X. Liu, Y . Zhang, H. Jing, L. Wang, and S. Zhao, “Ore image segmentation method using U-Net and Res unet convolutional networks,” RSC Advances , vol. 10, no. 16, pp. 9396–9406, 2020. [Online]. Available: https://xlink.rsc.org/?DOI=C9RA05877J

    work page 2020

  47. [47]

    Multidimensional characterization of particle morphology and mineralogical composition using ct data and r- vine copulas,

    O. Furat, T. Kirstein, T. Leißner, K. Bachmann, J. Gutzmer, U. A. Peuker, and V . Schmidt, “Multidimensional characterization of particle morphology and mineralogical composition using ct data and r- vine copulas,” Minerals Engineering , vol. 206, p. 108520, 2024. [Online]. Available: https://www.sciencedirect.com/science/article/pii/ S0892687523005344

    work page 2024

  48. [48]

    nnU-Net: a self-configuring method for deep learning-based biomedical image segmentation,

    F. Isensee, P. F. Jaeger, S. A. A. Kohl, J. Petersen, and K. H. Maier-Hein, “nnU-Net: a self-configuring method for deep learning-based biomedical image segmentation,” Nature Methods , vol. 18, no. 2, pp. 203–211, Feb. 2021. [Online]. Available: https://www.nature.com/articles/s41592-020-01008-z

    work page 2021

  49. [49]

    Unsupervised word sense disambiguation rivaling supervised methods,

    D. Yarowsky, “Unsupervised word sense disambiguation rivaling supervised methods,” in 33rd Annual Meeting of the Association for Computational Linguistics . Cambridge, Massachusetts, USA: Association for Computational Linguistics, Jun. 1995, pp. 189–196. [Online]. Available: https://aclanthology.org/P95-1026/

    work page 1995

  50. [50]

    Text Classification from Labeled and Unlabeled Documents using EM,

    K. Nigam, A. K. Mccallum, S. Thrun, and T. Mitchell, “Text Classification from Labeled and Unlabeled Documents using EM,” Machine Learning , vol. 39, no. 2/3, pp. 103–134, 2000. [Online]. Available: http://link.springer.com/10.1023/A:1007692713085

    doi:10.1023/a:1007692713085 2000

  51. [51]

    Self-training with noisy student improves imagenet classification,

    Q. Xie, M.-T. Luong, E. Hovy, and Q. V . Le, “Self-training with noisy student improves imagenet classification,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), June 2020

    work page 2020

  52. [52]

    Rethinking pre-training and self-training,

    B. Zoph, G. Ghiasi, T.-Y . Lin, Y . Cui, H. Liu, E. D. Cubuk, and Q. V . Le, “Rethinking pre-training and self-training,” in Proceedings of the 34th International Conference on Neural Information Processing Systems, ser. NIPS ’20. Red Hook, NY , USA: Curran Associates Inc., 2020

    work page 2020

  53. [53]

    Unsupervised data augmentation for consistency training,

    Q. Xie, Z. Dai, E. Hovy, M.-T. Luong, and Q. V . Le, “Unsupervised data augmentation for consistency training,” in Proceedings of the 34th International Conference on Neural Information Processing Systems, ser. NIPS ’20. Red Hook, NY , USA: Curran Associates Inc., 2020

    work page 2020

  54. [54]

    Self-training for end-to-end speech recognition,

    J. Kahn, A. Lee, and A. Hannun, “Self-training for end-to-end speech recognition,” in ICASSP 2020 - 2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) , 2020, pp. 7084– 7088

    work page 2020

  55. [55]

    Semi- supervised learning for network-based cardiac mr image segmentation,

    W. Bai, O. Oktay, M. Sinclair, H. Suzuki, M. Rajchl, G. Tarroni, B. Glocker, A. King, P. M. Matthews, and D. Rueckert, “Semi- supervised learning for network-based cardiac mr image segmentation,” in Medical Image Computing and Computer-Assisted Intervention - MICCAI 2017 , M. Descoteaux, L. Maier-Hein, A. Franz, P. Jannin, D. L. Collins, and S. Duchesne, ...

    work page 2017

  56. [56]

    Local contrastive loss with pseudo-label based self-training for semi-supervised medical image segmentation,

    K. Chaitanya, E. Erdil, N. Karani, and E. Konukoglu, “Local contrastive loss with pseudo-label based self-training for semi-supervised medical image segmentation,” Medical Image Analysis, vol. 87, p. 102792, 2023. [Online]. Available: https://www.sciencedirect.com/science/article/pii/ S1361841523000531

    work page 2023

  57. [57]

    Uncertainty-aware self-training for few-shot text classification,

    S. Mukherjee and A. H. Awadallah, “Uncertainty-aware self-training for few-shot text classification,” in Proceedings of the 34th International Conference on Neural Information Processing Systems , ser. NIPS ’20. Red Hook, NY , USA: Curran Associates Inc., 2020

    work page 2020

  58. [58]

    Anatomically-aware uncertainty for semi-supervised image segmentation,

    S. Adiga V ., J. Dolz, and H. Lombaert, “Anatomically-aware uncertainty for semi-supervised image segmentation,” Medical Image Analysis, vol. 91, p. 103011, 2024. [Online]. Available: https: //www.sciencedirect.com/science/article/pii/S1361841523002712

    work page 2024

  59. [59]

    Fixmatch: simplifying semi- supervised learning with consistency and confidence,

    K. Sohn, D. Berthelot, C.-L. Li, Z. Zhang, N. Carlini, E. D. Cubuk, A. Kurakin, H. Zhang, and C. Raffel, “Fixmatch: simplifying semi- supervised learning with consistency and confidence,” in Proceedings of the 34th International Conference on Neural Information Processing Systems, ser. NIPS ’20. Red Hook, NY , USA: Curran Associates Inc., 2020

    work page 2020

  60. [60]

    Flexmatch: boosting semi-supervised learning with curriculum pseudo labeling,

    B. Zhang, Y . Wang, W. Hou, H. Wu, J. Wang, M. Okumura, and T. Shi- nozaki, “Flexmatch: boosting semi-supervised learning with curriculum pseudo labeling,” in Proceedings of the 35th International Conference on Neural Information Processing Systems , ser. NIPS ’21. Red Hook, NY , USA: Curran Associates Inc., 2021

    work page 2021

  61. [61]

    Curriculum label- ing: Revisiting pseudo-labeling for semi-supervised learning,

    P. Cascante-Bonilla, F. Tan, Y . Qi, and V . Ordonez, “Curriculum label- ing: Revisiting pseudo-labeling for semi-supervised learning,” 2020

    work page 2020

  62. [62]

    NeTra: A toolbox for navigating large image databases,

    W.-Y . Ma and B. Manjunath, “NeTra: A toolbox for navigating large image databases,” Multimedia Systems , vol. 7, no. 3, pp. 184–198, May 1999. [Online]. Available: http://link.springer.com/10. 1007/s005300050121

    work page 1999

  63. [63]

    PyCUDA and PyOpenCL: A scripting-based approach to GPU run-time code generation,

    A. Kl ¨ockner, N. Pinto, Y . Lee, B. Catanzaro, P. Ivanov, and A. Fasih, “PyCUDA and PyOpenCL: A scripting-based approach to GPU run-time code generation,” Parallel Computing , vol. 38, no. 3, pp. 157–174, Mar. 2012. [Online]. Available: https://linkinghub.elsevier. com/retrieve/pii/S0167819111001281

    work page 2012

  64. [64]

    Sam 2: Segment anything in images and videos,

    N. Ravi, V . Gabeur, Y .-T. Hu, R. Hu, C. Ryali, T. Ma, H. Khedr, R. R ¨adle, C. Rolland, L. Gustafson, E. Mintun, J. Pan, K. V . Alwala, N. Carion, C.-Y . Wu, R. Girshick, P. Doll ´ar, and C. Feichtenhofer, “Sam 2: Segment anything in images and videos,” 2024. [Online]. Available: https://arxiv.org/abs/2408.00714

    Pith/arXiv arXiv 2024

  65. [65]

    A systematic study of the class imbalance problem in convolutional neural networks,

    M. Buda, A. Maki, and M. A. Mazurowski, “A systematic study of the class imbalance problem in convolutional neural networks,” Neural Networks , vol. 106, pp. 249–259, 2018. [Online]. Available: https://www.sciencedirect.com/science/article/pii/S0893608018302107

    work page 2018

  66. [66]

    An FFT-based technique for translation, rotation, and scale-invariant image registration,

    B. Reddy and B. Chatterji, “An FFT-based technique for translation, rotation, and scale-invariant image registration,” IEEE Transactions on Image Processing, vol. 5, no. 8, pp. 1266–1271, Aug. 1996. [Online]. Available: https://ieeexplore.ieee.org/document/506761/

    work page 1996

  67. [67]

    nnInteractive: Redefining 19 3d promptable segmentation,

    F. Isensee, M. Rokuss, L. Kr ¨amer, S. Dinkelacker, A. Ravindran, F. Stritzke, B. Hamm, T. Wald, M. Langenberg, C. Ulrich, J. Deissler, R. Floca, and K. Maier-Hein, “nnInteractive: Redefining 19 3d promptable segmentation,” 2025. [Online]. Available: https: //arxiv.org/abs/2503.08373

    Pith/arXiv arXiv 2025

  68. [68]

    Techniques in high-speed imaging and X-ray micro-computed tomography for characterisation of iron ore fragmentation

    A. Barron, Y . Zhang, N. Kandula, M. Shadwell, S. Bensley, T. Evans, L. O’Connor, M. Knackstedt, and N. Francois, “Techniques in high-speed imaging and X-ray micro-computed tomography for characterisation of iron ore fragmentation,” 2025, version Number: 1. [Online]. Available: https://arxiv.org/abs/2503.03163

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  69. [69]

    Space-filling x-ray source trajectories for efficient scanning in large-angle cone-beam computed tomography,

    A. M. Kingston, G. R. Myers, S. J. Latham, B. Recur, H. Li, and A. P. Sheppard, “Space-filling x-ray source trajectories for efficient scanning in large-angle cone-beam computed tomography,” IEEE Transactions on Computational Imaging, vol. 4, no. 3, pp. 447–458, 2018

    work page 2018

  70. [70]

    PyTorch: an imperative style, high- performance deep learning library,

    A. Paszke, S. Gross, F. Massa, A. Lerer, J. Bradbury, G. Chanan, T. Killeen, Z. Lin, N. Gimelshein, L. Antiga, A. Desmaison, A. K ¨opf, E. Yang, Z. DeVito, M. Raison, A. Tejani, S. Chilamkurthy, B. Steiner, L. Fang, J. Bai, and S. Chintala, “PyTorch: an imperative style, high- performance deep learning library,” in Proceedings of the 33rd Inter- national ...

    work page 2019

  71. [71]

    Is sam 2 better than sam in medical image segmentation?

    S. Sengupta, S. Chakrabarty, and R. Soni, “Is sam 2 better than sam in medical image segmentation?” in Medical Imaging 2025: Image Processing, O. Colliot and J. Mitra, Eds. SPIE, Apr. 2025, p. 97. [Online]. Available: http://dx.doi.org/10.1117/12.3047370

    work page doi:10.1117/12.3047370 2025

  72. [72]

    Chollet et al., “Keras,” https://keras.io, 2015

    F. Chollet et al., “Keras,” https://keras.io, 2015

    2015

  73. [73]

    TensorFlow: a system for large-scale machine learning,

    M. Abadi, P. Barham, J. Chen, Z. Chen, A. Davis, J. Dean, M. Devin, S. Ghemawat, G. Irving, M. Isard, M. Kudlur, J. Levenberg, R. Monga, S. Moore, D. G. Murray, B. Steiner, P. Tucker, V . Vasudevan, P. Warden, M. Wicke, Y . Yu, and X. Zheng, “TensorFlow: a system for large-scale machine learning,” in Proceedings of the 12th USENIX Conference on Operating ...

    work page 2016

  74. [74]

    3D U-Net: Learning Dense V olumetric Segmentation from Sparse Annotation,

    ¨O. C ¸ ic ¸ek, A. Abdulkadir, S. S. Lienkamp, T. Brox, and O. Ronneberger, “3D U-Net: Learning Dense V olumetric Segmentation from Sparse Annotation,” 2016, version Number: 1. [Online]. Available: https: //arxiv.org/abs/1606.06650

    Pith/arXiv arXiv 2016

  75. [75]

    Natural variability in bee brain size and symmetry revealed by micro-CT imaging and deep learning,

    P. D. L ¨osel, C. Monchanin, R. Lebrun, A. Jayme, J. J. Relle, J.-M. Devaud, V . Heuveline, and M. Lihoreau, “Natural variability in bee brain size and symmetry revealed by micro-CT imaging and deep learning,” PLOS Computational Biology , vol. 19, no. 10, p. e1011529, Oct. 2023. [Online]. Available: https://dx.plos.org/10.1371/journal.pcbi.1011529

    work page doi:10.1371/journal.pcbi.1011529 2023

  76. [76]

    Automated 3D cytoplasm segmentation in soft X-ray tomography,

    A. Erozan, P. D. L ¨osel, V . Heuveline, and V . Weinhardt, “Automated 3D cytoplasm segmentation in soft X-ray tomography,” iScience, vol. 27, no. 6, p. 109856, Jun. 2024. [Online]. Available: https: //linkinghub.elsevier.com/retrieve/pii/S2589004224010782

    work page 2024

  77. [77]

    HEDI: First- Time Clinical Application and Results of a Biomechanical Evaluation and Visualisation Tool for Incisional Hernia Repair,

    P. D. L ¨osel, J. J. Relle, S. V oß, R. Raschidi, R. Nessel, J. G ¨orich, M. O. Wielp¨utz, T. L ¨offler, V . Heuveline, and F. Kallinowski, “HEDI: First- Time Clinical Application and Results of a Biomechanical Evaluation and Visualisation Tool for Incisional Hernia Repair,” 2023, version Number: 2. [Online]. Available: https://arxiv.org/abs/2307.01502

    work page arXiv 2023

  78. [78]

    Synchrotron-based 3D X-ray computed tomography reveals root system architecture: Plastic responses to phosphorus placement,

    F. J. Van Der Bom, P. D. L ¨osel, A. M. Kingston, Z. H. Weng, N. S. Raymond, A. Maksimenko, D. Hausermann, M. J. Bell, C. L. Doolette, E. Lombi, and P. M. Kopittke, “Synchrotron-based 3D X-ray computed tomography reveals root system architecture: Plastic responses to phosphorus placement,” Geoderma, vol. 457, p. 117299, May 2025. [Online]. Available: http...

    work page 2025

  79. [79]

    High-throughput phenomics of global ant biodiversity,

    J. Katzke, F. Hita Garcia, P. D. L ¨osel, F. Azuma, T. Farag´o, L. Aibekova, A. Casadei-Ferreira, S. Gautam, A. Richter, E. Toulkeridou, S. Bremer, E. Hamann, J. Hein, J. Odar, C. Sarkar, M. Zuber, J. J. Boomsma, R. M. Feitosa, L. Schrader, G. Zhang, S. Cs ˝osz, M. Dong, O. Evangelista, G. Fischer, B. L. Fisher, J. Florez Fernandez, The GAGA Consortium, F...

    work page 2025

  80. [80]

    3D Slicer: A Platform for Subject-Specific Image Analysis, Visualization, and Clinical Support,

    R. Kikinis, S. D. Pieper, and K. G. V osburgh, “3D Slicer: A Platform for Subject-Specific Image Analysis, Visualization, and Clinical Support,” in Intraoperative Imaging and Image-Guided Therapy , F. A. Jolesz, Ed. New York, NY: Springer New York, 2014, pp. 277–289. [Online]. Available: http://link.springer.com/10.1007/978-1-4614-7657-3 19

    work page doi:10.1007/978-1-4614-7657-3 2014

Showing first 80 references.

This paper was first reviewed by deepseek-v4-flash on August 5, 2026.