Pith. sign in

REVIEW 3 major objections 5 minor 45 references

JanusNet: Hierarchical Slice-Block Shuffle and Displacement for Semi-Supervised 3D Multi-Organ Segmentation

T0 review · 3 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read The paper claims that replacing cube-level mixing with an anatomy-preserving, same-index slice-block shuffle—plus a confidence-guided in-slice displacement—raises semi-supervised 3D multi-organ segmentation, e.g., +4% Dice on Synapse with 2

desk verdict A plausible, practical augmentation that deserves a serious referee, but the central alignment claim is untested and the manuscript as provided is incomplete. read the letter →

arxiv 2508.03997 v1 pith:XDEURCLM submitted 2025-08-06 cs.CV

Zheng Zhang , Tianzhuzi Tan , Guanchun Yin , Bo Zhang , Xiuzhuang Zhou This is my paper
classification cs.CV
keywords semi-supervisedlearning3Dmedicalimagesegmentationmulti-organdataaugmentationslice-blockshuffleconfidence-guideddisplacementteacher-studentanatomicalcontinuity
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

Semi-supervised 3D multi-organ segmentation needs to exploit unlabeled CT volumes, and data augmentation is a key lever. JanusNet claims that the common trick of mixing random 3D cubes throws away exactly the information that matters—the smooth variation of organ shape and position along a volume's axes. So it shuffles aligned slice-blocks across labeled and unlabeled volumes along one randomly chosen axis, preserving in-plane anatomy, and then uses a teacher's confidence to move reliable patches into unreliable spots, focusing the model on hard regions. On the Synapse benchmark at 20% label ratio this produces 72.67% average Dice, about 4 points above the strongest prior method; on AMOS at 5% it produces 63.99%.

What carries the argument

Slice-Block Shuffle (SBS): partition each volume along a randomly chosen orthogonal axis and shuffle same-index slice blocks across a batch's labeled and unlabeled volumes, so the orthogonal planes retain anatomical continuity. Confidence-Guided Displacement (CGD): within each produced slice, replace low-reliability blocks (from the teacher's softmax confidence) with high-reliability blocks to emphasize hard-to-segment regions. Together they form a two-stage, layer-aware augmentation inside a teacher-student loop with an EMA teacher, pseudo-labels, and a displacement loss.

What would settle it

Run the same 20%-labeled Synapse protocol on a multi-center CT dataset where axial slice indices are known to vary across patients (e.g., different start positions). If JanusNet's average Dice drops to or below GA-MagicNet's, the index-alignment assumption is the load-bearing part. A direct control is to repeat JanusNet with index alignment replaced by random slice-block mixing: the gain should vanish.

Watch

Extended reading notes

Core claim

The paper's central claim is that axis-aligned, slice-block-level mixing is a better perturbation for semi-supervised 3D multi-organ segmentation than cube-level mixing. JanusNet picks one of the three volume axes, splits labeled and unlabeled volumes into blocks along that axis, and shuffles same-index blocks across volumes, preserving the anatomical layout on the planes perpendicular to the chosen axis. A second step, Confidence-Guided Displacement, replaces low-confidence regions inside each slice with high-confidence counterparts from the same slice, using the teacher's predictions to home in on hard organs. In the reported experiments, the two-stage augmentation lifts average Dice on 20

Load-bearing premise

The strongest assumption is that the same slice index across different patient volumes refers to comparable anatomy, so shuffling same-index slice blocks mixes consistent content; unregistered CT scans with shifted table positions could break this.

Editorial extensions

If this is right

  • If the reported comparisons hold, JanusNet's two augmentation steps can be dropped into existing teacher-student semi-supervised segmentation frameworks with minimal code changes and still yield consistent gains.
  • The per-class results indicate that the method helps most where cube mixing hurts most: small or elongated organs. On Synapse, stomach, pancreas, and spleen improve by 12.3, 11.6, and 5.4 points; on AMOS, esophagus, stomach, and duodenum improve by 9.2, 7.4, and 6.3 points.
  • The ablations show a well-defined optimum in perturbation strength (block thickness p=16, displacement weight 0.25, Top-K=2): too little mixing leaves hard regions untouched, too much erodes the anatomical prior and degrades Dice, giving practical guidance for applying the method to new datasets.
  • Because the gains are measured on both 20%-labeled Synapse and 5%-labeled AMOS, the method appears to stay effective under severe annotation scarcity, the regime where semi-supervised segmentation is most needed.

Reading between the lines

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

  • The method's reliance on same-index slice correspondence is the main untested risk. In multi-center CT, axial slice index does not necessarily align to anatomy, so an obvious next experiment is to test JanusNet on scans with different table positions or to prepend a simple anatomical-level normalization; the paper does not do this.
  • Because the augmentation only needs intensity volumes and pseudo-labels, it should transfer to 3D MRI and other modalities where slice-index correspondence is often more reliable, though the authors only evaluate on CT.
  • The success of axis-aligned mixing hints that the anatomical prior is what matters more than the specific mixing mechanism. A cleaner test would compare JanusNet against cube mixing under identical compute and schedule, isolating alignment rather than block size or loss weights as the source of the gain.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 5 minor

Summary. The paper proposes JanusNet, a data-augmentation framework for semi-supervised 3D multi-organ segmentation. It introduces two stages built on a teacher–student setup: Slice-Block Shuffle (SBS), which mixes blocks of slices at the same index across volumes along a randomly chosen axis, and Confidence-Guided Displacement (CGD), which replaces low-confidence blocks with high-confidence counterparts within a slice. The claimed contribution is that this axis-aligned, slice-block-level perturbation preserves anatomical continuity better than cube-level mixing, improving learning on small and hard-to-segment organs. Experiments on Synapse (20% labeled) and AMOS (5% labeled) report state-of-the-art average Dice scores, with component ablations on Synapse showing stepwise gains for SBS and CGD over a weak–strong augmentation baseline.

Significance. If the reported results hold, JanusNet would be a simple, plug-and-play augmentation that improves several semi-supervised 3D segmentation baselines, which is practically valuable. The paper's component ablations (Table 3) show consistent positive contributions from both proposed steps, and the gains on small organs such as the stomach and pancreas are directionally encouraging. The method is also agnostic to the backbone and thus potentially easy to adopt. However, the central claim rests on the anatomical alignment being beneficial; the manuscript does not isolate this mechanism, and the main quantitative comparisons lack statistical support. These gaps currently limit the strength of the conclusions and the confidence one can place in the reported SOTA numbers.

major comments (3)
  1. [§3.1, Fig. 1, Tables 3] The central mechanism of Slice-Block Shuffle is that aligned same-index slice blocks across volumes preserve anatomical context. However, the paper never tests whether this alignment, rather than generic block mixing, causes the +2.84% gain over the weak-strong baseline (Table 3). A control experiment mixing blocks at random indices or random offsets is necessary; otherwise the method could be another variant of CutMix/MixUp with no unique anatomical contribution. Additionally, the paper does not state how volumes with differing depth, spacing, or scan range are resampled or padded to form same-index blocks, which is essential to interpret the alignment claim.
  2. [Tables 1 and 2] The main comparisons against prior methods report only point estimates of Avg. Dice and ASD, with no standard deviations or significance tests. For example, JanusNet's 63.99 vs. GA-MagicNet's 63.51 on AMOS (Table 2) is a difference of 0.48 percentage points, which may be within run-to-run variability. The abstract's phrase 'significantly surpasses' is therefore not supported by the evidence as presented. The authors should report means and standard deviations over multiple runs (as is done for ablations in Table 3) and, where appropriate, significance tests or at least paired per-class comparisons.
  3. [Tables 3–6 versus Table 1] The default hyperparameters p=16, lambda_disp=0.25, and Top-K=2 are selected by ablations on the same Synapse 20% labeled benchmark that is used for the headline SOTA comparison (Table 1). Such test-set-based selection can overstate gains. The authors should either use a separate validation split for hyperparameter selection, report results for a sensible range of hyperparameters, or provide a nested evaluation to demonstrate that the chosen settings generalize. Without this, the reported Synapse gain may be optimistic.
minor comments (5)
  1. [Abstract] The abstract describes the setting as 'weakly supervised' while the paper is semi-supervised; please use consistent terminology throughout.
  2. [Tables 4–6] The number of runs used to compute mean ± std is not stated; please specify the number of independent repetitions.
  3. [References] The text cites 'MagicNet(Qi, Wu, and Chan 2024)' but the reference list entry is incomplete; also check that all cited works (e.g., Wang and Li 2023, Wu et al. 2022) appear in the references.
  4. [Figures 1–2] The figures are difficult to read in the submitted version, especially the small inset labels in Fig. 1 and the data flow annotations in Fig. 2. Please provide higher-resolution versions and enlarge key sub-captions.
  5. [§3.2/3.3] The method description for the two proposed steps is sparse; pseudocode or a more detailed algorithmic description would improve reproducibility, particularly for CGD's Top-K selection and the exact formation of displaced targets.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: JanusNet's claims are empirical, tested against external datasets and prior methods; no derived quantity reduces to a fitted parameter or self-citation.

full rationale

The augmentation framework (Slice-Block Shuffle and Confidence-Guided Displacement) is defined independently of the reported metric. The 'prediction' is a segmentation Dice score on held-out Synapse/AMOS data, obtained by training a teacher-student network with the proposed augmentations and comparing to published baselines. No equation in the paper equates the reported gain to the augmentation by construction; the gain is an experimental outcome. Hyperparameters p, lambda_disp, and Top-K are tuned via ablations and then fixed; they are standard hyperparameters, not quantities derived from the target metric and then re-reported as predictions. The only self-referential element is the use of the EMA teacher's pseudo-labels in the Confidence-Guided Displacement step; this is a standard consistency-training loop (teacher produces pseudo-labels, student is trained against them), not a circular derivation of the reported improvement. The reference to Zhang et al. 2025 (SKCDF) is used as a comparison baseline, not as load-bearing support for JanusNet's mechanism. The possible concern that same-index slice alignment across unregistered CT volumes is not validated by a random-offset control is an experimental/validity question, not a circularity: the method's success or failure does not reduce to an assumed definition. Therefore no circular step meeting the stated criteria is present.

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

The central claim is empirical and does not rest on a formal derivation. The free parameters are benchmark-tuned design choices (p, lambda_disp, Top-K) selected on the same Synapse setting whose result is reported. The axioms are domain assumptions about anatomy, slice-index correspondence, and pseudo-label reliability; none is independently verified in the paper.

free parameters (4)
  • slice-block thickness p = 16
    Selected by ablation on 20% Synapse (Table 4) to maximize average Dice; affects how much structure is preserved versus perturbed.
  • displacement loss weight lambda_disp = 0.25
    Selected by ablation on 20% Synapse (Table 5); larger values degrade Dice monotonically.
  • Top-K (number of hard blocks displaced) = 2
    Selected by ablation on 20% Synapse (Table 6); K=1 under-covers difficult regions and K>=3 over-perturbs.
  • overall loss weights alpha and beta
    The objective in Equation 1 defines beta = alpha * lambda_disp, but alpha is not specified in the provided text; this affects the balance of supervised, pseudo-label, and displacement losses.
assumptions (4)
  • domain assumption Preserving anatomical continuity on planes orthogonal to the perturbed axis improves semi-supervised segmentation.
    Underlies the design of Slice-Block Shuffle; stated in Section 1 and Figure 1.
  • domain assumption Same-index slice blocks across different patient volumes correspond to comparable anatomical levels.
    Required for the 'aligned shuffling of same-index slice blocks' in Section 3.2 to produce semantically plausible mixed volumes; no registration or normalization is described.
  • domain assumption EMA teacher confidence is a reliable proxy for segmentation difficulty and correctness.
    Confidence-Guided Displacement uses prediction reliability to select unreliable blocks; assumption stated in Section 3.3.
  • standard math The teacher-student framework with Dice loss and EMA is a valid base for semi-supervised segmentation.
    Adopted from prior work (Tarvainen and Valpola 2017; Yu et al. 2019), used throughout the experiments.

how reviews work

0 comments
Cite this review

Pith. "Pith review of JanusNet: Hierarchical Slice-Block Shuffle and Displacement for Semi-Supervised 3D Multi-Organ Segmentation." pith.science (2026). https://pith.science/paper/XDEURCLM

@misc{pith2026250803997,
  author       = {Pith},
  title        = {Pith review of: JanusNet: Hierarchical Slice-Block Shuffle and Displacement for Semi-Supervised 3D Multi-Organ Segmentation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/XDEURCLM}},
  note         = {Machine review of arXiv:2508.03997}
}
read the original abstract

Limited by the scarcity of training samples and annotations, weakly supervised medical image segmentation often employs data augmentation to increase data diversity, while randomly mixing volumetric blocks has demonstrated strong performance. However, this approach disrupts the inherent anatomical continuity of 3D medical images along orthogonal axes, leading to severe structural inconsistencies and insufficient training in challenging regions, such as small-sized organs, etc. To better comply with and utilize human anatomical information, we propose JanusNet}, a data augmentation framework for 3D medical data that globally models anatomical continuity while locally focusing on hard-to-segment regions. Specifically, our Slice-Block Shuffle step performs aligned shuffling of same-index slice blocks across volumes along a random axis, while preserving the anatomical context on planes perpendicular to the perturbation axis. Concurrently, the Confidence-Guided Displacement step uses prediction reliability to replace blocks within each slice, amplifying signals from difficult areas. This dual-stage, axis-aligned framework is plug-and-play, requiring minimal code changes for most teacher-student schemes. Extensive experiments on the Synapse and AMOS datasets demonstrate that JanusNet significantly surpasses state-of-the-art methods, achieving, for instance, a 4% DSC gain on the Synapse dataset with only 20% labeled data.

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

45 extracted references · 40 canonical work pages

  1. [1]

    E.; and McGuinness, K

    Arazo, E.; Ortego, D.; Albert, P.; O’Connor, N. E.; and McGuinness, K. 2020. Pseudo-labeling and confirmation bias in deep semi-supervised learning. In IJCNN'20, 1--8

    work page 2020

  2. [2]

    M.; and Rueckert, D

    Bai, W.; Oktay, O.; Sinclair, M.; Suzuki, H.; Rajchl, M.; Tarroni, G.; Glocker, B.; King, A.; Matthews, P. M.; and Rueckert, D. 2017. Semi-supervised learning for network-based cardiac MR image segmentation. In MICCAI'17, 253--260

    work page 2017

  3. [3]

    Bai, Y.; Chen, D.; Li, Q.; Shen, W.; and Wang, Y. 2023. Bidirectional copy-paste for semi-supervised medical image segmentation. In CVPR'23, 11514--11524

    work page 2023

  4. [4]

    Basak, H.; Ghosal, S.; and Sarkar, R. 2022. Addressing class imbalance in semi-supervised image segmentation: A study on cardiac mri. In MICCAI'22, 224--233

    work page 2022

  5. [5]

    D.; Kurakin, A.; Sohn, K.; Zhang, H.; and Raffel, C

    Berthelot, D.; Carlini, N.; Cubuk, E. D.; Kurakin, A.; Sohn, K.; Zhang, H.; and Raffel, C. 2019 a . Remixmatch: Semi-supervised learning with distribution alignment and augmentation anchoring. arXiv preprint arXiv:1911.09785

    arXiv 2019

  6. [6]

    Berthelot, D.; Carlini, N.; Goodfellow, I.; Papernot, N.; Oliver, A.; and Raffel, C. A. 2019 b . Mixmatch: A holistic approach to semi-supervised learning. NeurIPS'19, 32

    work page 2019

  7. [7]

    Cai, H.; Li, S.; Qi, L.; Yu, Q.; Shi, Y.; and Gao, Y. 2023. Orthogonal annotation benefits barely-supervised medical image segmentation. In CVPR'23, 3302--3311

    work page 2023

  8. [8]

    Chen, B.; Jiang, J.; Wang, X.; Wan, P.; Wang, J.; and Long, M. 2022 a . Debiased self-training for semi-supervised learning. NeurIPS'22, 35: 32424--32437

    work page 2022

Show all 45 references
  1. [9]

    Chen, D.; Bai, Y.; Shen, W.; Li, Q.; Yu, L.; and Wang, Y. 2023. Magicnet: Semi-supervised multi-organ segmentation via magic-cube partition and recovery. In CVPR'23, 23869--23878

    2023

  2. [10]

    Chen, H.; Fan, Y.; Wang, Y.; Wang, J.; Schiele, B.; Xie, X.; Savvides, M.; and Raj, B. 2022 b . An embarrassingly simple baseline for imbalanced semi-supervised learning. arXiv preprint arXiv:2211.11086

    2022 arXiv

  3. [11]

    Chen, J.; Mei, J.; Li, X.; Lu, Y.; Yu, Q.; Wei, Q.; Luo, X.; Xie, Y.; Adeli, E.; Wang, Y.; et al. 2024. TransUNet: Rethinking the U-Net architecture design for medical image segmentation through the lens of transformers. Medical Image Analysis, 97: 103280

    2024

  4. [12]

    Chen, X.; Yuan, Y.; Zeng, G.; and Wang, J. 2021. Semi-supervised semantic segmentation with cross pseudo supervision. In CVPR'21, 2613--2622

    2021

  5. [13]

    DeVries, T.; and Taylor, G. W. 2017. Improved regularization of convolutional neural networks with cutout. arXiv preprint arXiv:1708.04552

    2017 arXiv

  6. [14]

    Du, J.; Zhang, X.; Liu, P.; and Wang, T. 2023. Coarse-refined consistency learning using pixel-level features for semi-supervised medical image segmentation. IEEE Journal of Biomedical and Health Informatics, 27(8): 3970--3981

    2023

  7. [15]

    Guo, L.-Z.; and Li, Y.-F. 2022. Class-imbalanced semi-supervised learning with adaptive thresholding. In ICML'22, 8082--8094

    2022

  8. [16]

    He, J.; Cai, C.; Li, Q.; and Ma, A. J. 2024. Pair shuffle consistency for semi-supervised medical image segmentation. In MICCAI'24, 489--499

    2024

  9. [17]

    Hu, S.; Liao, Z.; and Xia, Y. 2022. Boundary-aware network for abdominal multi-organ segmentation. arXiv preprint arXiv:2208.13774

    2022 arXiv

  10. [18]

    Huang, W.; Chen, C.; Xiong, Z.; Zhang, Y.; Chen, X.; Sun, X.; and Wu, F. 2022. Semi-supervised neuron segmentation via reinforced consistency learning. IEEE Transactions on Medical Imaging, 41(11): 3016--3028

    2022

  11. [19]

    F.; Kohl, S

    Isensee, F.; Jaeger, P. F.; Kohl, S. A.; Petersen, J.; and Maier-Hein, K. H. 2021. nnU-Net: a self-configuring method for deep learning-based biomedical image segmentation. Nature methods, 18(2): 203--211

    2021

  12. [20]

    Kervadec, H.; Bouchtiba, J.; Desrosiers, C.; Granger, E.; Dolz, J.; and Ayed, I. B. 2019. Boundary loss for highly unbalanced segmentation. In MIDL'19, 285--296

    2019

  13. [21]

    Li, S.; Zhang, C.; and He, X. 2020. Shape-aware semi-supervised 3D semantic segmentation for medical images. In MICCAI'20, 552--561

    2020

  14. [22]

    Lin, Y.; Yao, H.; Li, Z.; Zheng, G.; and Li, X. 2022. Calibrating label distribution for class-imbalanced barely-supervised knee segmentation. In MICCAI'22, 109--118

    2022

  15. [23]

    Luo, X.; Chen, J.; Song, T.; and Wang, G. 2021 a . Semi-supervised medical image segmentation through dual-task consistency. In AAAI'21, 8801--8809

    2021

  16. [24]

    Luo, X.; Liao, W.; Chen, J.; Song, T.; Chen, Y.; Zhang, S.; Chen, N.; Wang, G.; and Zhang, S. 2021 b . Efficient semi-supervised gross target volume of nasopharyngeal carcinoma segmentation via uncertainty rectified pyramid consistency. In MICCAI'21, 318--329

    2021

  17. [25]

    Milletari, F.; Navab, N.; and Ahmadi, S.-A. 2016. V-net: Fully convolutional neural networks for volumetric medical image segmentation. In 3DV'16, 565--571

    2016

  18. [26]

    Qi, W.; Wu, J.; and Chan, S. 2024. Gradient-aware for class-imbalanced semi-supervised medical image segmentation. In ECCV'24, 473--490

    2024

  19. [27]

    Ronneberger, O.; Fischer, P.; and Brox, T. 2015. U-net: Convolutional networks for biomedical image segmentation. In MICCAI'15, 234--241

    2015

  20. [28]

    C.; Sekuboyina, A.; Ezhov, I.; Unger, A.; Zhylka, A.; Pluim, J

    Shit, S.; Paetzold, J. C.; Sekuboyina, A.; Ezhov, I.; Unger, A.; Zhylka, A.; Pluim, J. P.; Bauer, U.; and Menze, B. H. 2021. clDice-a novel topology-preserving loss function for tubular structure segmentation. In CVPR'21, 16560--16569

    2021

  21. [29]

    A.; Cubuk, E

    Sohn, K.; Berthelot, D.; Carlini, N.; Zhang, Z.; Zhang, H.; Raffel, C. A.; Cubuk, E. D.; Kurakin, A.; and Li, C.-L. 2020. Fixmatch: Simplifying semi-supervised learning with consistency and confidence. NeurIPS’20, 33: 596--608

    2020

  22. [30]

    N.; Wu, Z.; and Ding, X

    Tajbakhsh, N.; Jeyaseelan, L.; Li, Q.; Chiang, J. N.; Wu, Z.; and Ding, X. 2020. Embracing imperfect datasets: A review of deep learning solutions for medical image segmentation. Medical Image Analysis, 63: 101693

    2020

  23. [31]

    Tarvainen, A.; and Valpola, H. 2017. Mean teachers are better role models: Weight-averaged consistency targets improve semi-supervised deep learning results. NeurIPS’17, 30

    2017

  24. [32]

    Verma, V.; Kawaguchi, K.; Lamb, A.; Kannala, J.; Solin, A.; Bengio, Y.; and Lopez-Paz, D. 2022. Interpolation consistency training for semi-supervised learning. Neural Networks, 145: 90--106

    2022

  25. [33]

    Wang, F.; Zheng, K.; Lu, L.; Xiao, J.; Wu, M.; and Miao, S. 2021. Automatic vertebra localization and identification in CT by spine rectification and anatomically-constrained optimization. In CVPR'21, 5280--5288

    2021

  26. [34]

    Wang, H.; and Li, X. 2023. Towards generic semi-supervised framework for volumetric medical image segmentation. NeurIPS’23, 36

    2023

  27. [35]

    Wang, X.; Wu, Z.; Lian, L.; and Yu, S. X. 2022. Debiased learning from naturally imbalanced pseudo-labels. In CVPR'22, 14647--14657

    2022

  28. [36]

    K.; and Yuille, A

    Wang, Y.; Wei, X.; Liu, F.; Chen, J.; Zhou, Y.; Shen, W.; Fishman, E. K.; and Yuille, A. L. 2020. Deep distance transform for tubular structure segmentation in ct scans. In CVPR'20, 3833--3842

    2020

  29. [37]

    Wei, C.; Sohn, K.; Mellina, C.; Yuille, A.; and Yang, F. 2021. Crest: A class-rebalancing self-training framework for imbalanced semi-supervised learning. In CVPR'21, 10857--10866

    2021

  30. [38]

    Wu, Y.; Wu, Z.; Wu, Q.; Ge, Z.; and Cai, J. 2022. Exploring smoothness and class-separation for semi-supervised medical image segmentation. In MICCAI'22, 34--43

    2022

  31. [39]

    Yang, S. 2023. A review of research and development of semi-supervised learning strategies for medical image processing. EAI Endorsed Transactions on e-Learning, 9

    2023

  32. [40]

    Yu, L.; Wang, S.; Li, X.; Fu, C.-W.; and Heng, P.-A. 2019. Uncertainty-aware self-ensembling model for semi-supervised 3D left atrium segmentation. In MICCAI'19, 605--613

    2019

  33. [41]

    J.; Chun, S.; Choe, J.; and Yoo, Y

    Yun, S.; Han, D.; Oh, S. J.; Chun, S.; Choe, J.; and Yoo, Y. 2019. Cutmix: Regularization strategy to train strong classifiers with localizable features. In ICCV'19, 6023--6032

    2019

  34. [42]

    N.; and Lopez-Paz, D

    Zhang, H.; Cisse, M.; Dauphin, Y. N.; and Lopez-Paz, D. 2017. mixup: Beyond empirical risk minimization. arXiv preprint arXiv:1710.09412

    2017 arXiv

  35. [43]

    Zhang, Z.; Yin, G.; Zhang, B.; Liu, W.; Zhou, X.; and Wang, W. 2025. A Semantic Knowledge Complementarity based Decoupling Framework for Semi-supervised Class-imbalanced Medical Image Segmentation. In CVPR'25, 25940--25949

    2025

  36. [44]

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

    ENTRY address archivePrefix author booktitle chapter edition editor eid eprint howpublished institution isbn journal key month note number organization pages publisher school series title type volume year label extra.label sort.label short.list INTEGERS output.state before.all...

  37. [45]

    write newline

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

Pith tools

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