Pith. sign in

REVIEW 5 major objections 8 minor 28 references

TransMedSeg: A Transferable Semantic Framework for Semi-Supervised Medical Image Segmentation

T0 review · 5 major / 8 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read TransMedSeg adds a transferable semantic augmentation loss to a teacher-student segmentation framework and reports state-of-the-art semi-supervised results on cardiac and pancreas benchmarks.

desk verdict A plausible ISDA-style augmentation for SSMIS, but the theoretical derivation and empirical support are both weaker than claimed. read the letter →

arxiv 2505.14753 v1 pith:GFZPXD7J submitted 2025-05-20 eess.IV cs.AIcs.CV

Mengzhu Wang , Jiao Li , Shanshan Wang , Long Lan , Huibin Tan , Liang Yang , Guoli Yang This is my paper · ORCID
classification eess.IVcs.AIcs.CV
keywords semi-supervisedmedicalimagesegmentationtransferablesemanticaugmentationteacher-studentframeworkdomainshiftclass-conditionalcovarianceupper-boundlosspseudo-labeling
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The paper claims that its framework TransMedSeg, which adds a Transferable Semantic Augmentation (TSA) loss to a graph-based teacher-student baseline, consistently outperforms existing semi-supervised medical image segmentation methods on the ACDC, Pancreas-NIH, and LA benchmarks, reaching best Dice scores of 89.96%, 83.06%, and 89.62%. The key idea is to model the feature-space gap between labeled and unlabeled data as a class-conditional Gaussian perturbation whose mean is the teacher-student difference and whose covariance is the teacher's intra-class covariance, then minimize an upper bound of the expected cross-entropy under that perturbation. This makes the augmentation implicit, so no extra features are generated and the computational overhead is negligible. A sympathetic reader would care because the method targets two practical problems in medical imaging: domain shift across scanners or populations, and over-reliance on scarce labeled annotations.

What carries the argument

The load-bearing object is the Transferable Semantic Augmentation (TSA) module and its upper-bound surrogate loss. TSA samples feature perturbations from a class-conditional Gaussian δ ∼ N(α∆μ^c, αΣ_t^c), where ∆μ^c = μ_t^c − μ_s^c is the inter-domain mean shift and Σ_t^c is the teacher's EMA-estimated intra-class covariance; the perturbation is applied to student features and the expected cross-entropy is replaced by the closed-form upper bound L_tsa = (1/n_s) Σ_i log Σ_{c=1}^C E[exp(Δw_c^T(f_i+δ)+Δb_c)]. This machinery lets an infinite number of augmented views be accounted for implicitly, preserving anatomical structure through the covariance while avoiding explicit sampling and large memory cost.

What would settle it

Take the unlabeled set of ACDC, deliberately corrupt a fixed fraction (say 20%) of the pseudo-labels by swapping the right-ventricle and left-ventricle classes, and recompute the teacher EMA covariance; if TransMedSeg's Dice on the validation split does not degrade substantially relative to the same corruption applied without the TSA loss, then the covariance estimate is not actually the load-bearing channel the paper claims.

Watch

Extended reading notes

Core claim

On its own terms, TransMedSeg establishes that cross-domain feature alignment for semi-supervised segmentation does not require explicit data generation. For each anatomical class, the student's source statistics (μ_s^c, Σ_s^c) and the teacher's target statistics (μ_t^c, Σ_t^c), the latter updated by EMA over pseudo-labeled unlabeled features, are combined into a perturbation δ ∼ N(α(μ_t^c − μ_s^c), αΣ_t^c). The expected cross-entropy of a classifier applied to source features perturbed this way is shown, via Jensen's inequality and the strong law of large numbers, to be bounded above by an explicit log-sum-exp loss, which the paper minimizes as L_tsa. Added to the GraphCL loss, this surrogate improves Dice and boundary metrics (95HD, ASD) across cardiac MRI, pancreas CT, and left-atrium MRI, under label ratios from 5% to 20%. The authors therefore position TSA as a plug-in transferable semantic augmentation for other SSMIS methods.

Load-bearing premise

The whole benefit rests on the teacher's class-conditional covariance Σ_t^c, estimated by EMA over pseudo-labeled unlabeled features, being a faithful model of the true target class distribution; if wrong-class pixels pollute that estimate, the Gaussian perturbation will amplify mislabeled anatomical variation and can hurt the student.

Editorial extensions

If this is right

  • If the claim is right, semi-supervised segmentation frameworks can inherit transferable semantic alignment as a drop-in loss term without changing their architecture.
  • The reported gains on 95HD and ASD indicate that the method is particularly useful when clinical workflows care about boundary fidelity, not just overlap.
  • The method's effectiveness at 5% labeled data on ACDC and LA suggests it can operate in the annotation-scarce regime that real clinical settings often face.
  • Because the loss is an upper bound rather than an exact expectation, it can be tuned by the weight β (set to 0.4) to trade off transfer strength against source fidelity.

Reading between the lines

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

  • One testable extension of the same mechanism is to apply the TSA loss to a deliberately shifted test distribution (e.g., a second scanner with different contrast), to separate genuine domain alignment from generic feature smoothing; the paper reports only intra-dataset splits.
  • The Gaussian assumption for class-conditional features is a strong simplification; in classes with highly non-elliptical or multi-modal feature geometry, the covariance-based bound may loosen, and the method might benefit from a heavier-tailed perturbation family.
  • The theoretical guarantee is an upper bound on the expected loss, not an equivalence; practitioners should expect the bound to be loose in early training, and the EMA covariance to lag behind the student, so the optimal α and β likely depend on the label ratio.
  • A natural next experiment would use the same L_tsa to bridge labeled and unlabeled data from different institutions (e.g., different MR vendors), where the domain shift is known to be larger than within-dataset split.
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

5 major / 8 minor

Summary. TransMedSeg is a semi-supervised medical image segmentation method built on the authors' GraphCL framework. It adds a Transferable Semantic Augmentation (TSA) loss that models domain shift by sampling Gaussian perturbations in feature space, with class-conditional means and covariances estimated from labeled source features and EMA-updated pseudo-labeled target features. The authors claim an upper-bound formulation of the expected cross-entropy loss over infinite augmentations, making the method theoretically grounded and computationally cheap. They report experiments on ACDC, Pancreas-NIH, and LA with small labeled fractions (3–12 scans), comparing against several SSL baselines and report improved Dice, Jaccard, 95HD, and ASD metrics, plus ablations and t-SNE visualizations.

Significance. If the theoretical derivation were correct and the empirical gains were reproducible with uncertainty estimates, the paper would offer a useful plug-in regularizer for teacher-student semi-supervised medical image segmentation; the ablation study in Table 4 and the t-SNE visualizations are helpful evidence that the loss can affect feature geometry. However, as printed the central 'theoretically grounded' claim is invalid, the sign of the domain shift is inconsistent, and the experimental evidence is single-run and mostly compared against the authors' own prior baseline. The cross-domain/transferable claim in the title and abstract is not tested by any cross-site, cross-scanner, or cross-modality experiment. The contribution is best understood as an application of ISDA-style implicit augmentation to a tailored teacher-student SSL setting, which could be of interest if properly framed and validated.

major comments (5)
  1. [Section 2.2, Eqs. (6)–(8)] The derivation of the upper-bound loss is not valid as written. The cross-entropy in Eq. (6) equals log(sum_c exp((w_c - w_{y_i})^T f + (b_c - b_{y_i}))) after subtracting the numerator from the denominator, so the correct definitions are Δw_c = w_c - w_{y_i} and Δb_c = b_c - b_{y_i}. The text instead defines Δw_c ≡ w_{y_i}, which removes the class index c that the outer sum over c requires and does not correspond to Eq. (6). Moreover, Eq. (8) still contains E[exp(...)]; no Gaussian moment-generating function is evaluated, so the claimed closed-form upper-bound loss is not actually delivered. The central 'theoretically grounded' assertion is therefore unsupported and needs to be corrected.
  2. [Section 2.1 vs. Section 2.2, Eq. (3)] The direction of the domain shift is defined with opposite signs in the two sections. In Section 2.1, the text states Δμ^c = μ_s^c - μ_t^c, while Eq. (3) defines Δμ^c = μ_t^c - μ_s^c. Because the augmentation is f + δ with δ ∼ N(αΔμ, αΣ_t), this sign determines whether source features are moved toward or away from the target statistics. As printed, the method and its intended semantic alignment are ambiguous; the authors must choose one convention and use it consistently throughout.
  3. [Section 3.2, Tables 1–4] The central empirical claim of 'consistently outperforms state-of-the-art' is not supported by the reported numbers. All results are single-run point estimates with no standard deviations, repeated-seed trials, or significance tests. The improvements over GraphCL are small in several settings (e.g., LA 10% Dice 89.62 vs. 89.49; ACDC 5% Dice 88.14 vs. 87.96; Pancreas Dice 83.06 vs. 82.78), and the claim of consistent superiority cannot be assessed without variance estimates. In addition, GraphCL (Ref. [17]) is the authors' own prior work, and TransMedSeg is GraphCL plus L_tsa (Eq. (9)); to support the SOTA claim, the authors should compare against an independently implemented strong baseline and report uncertainty.
  4. [Abstract and Section 3.1] The paper claims transferability 'across different clinical domains and imaging modalities' and describes TSA as cross-domain distribution matching, but all experiments are standard semi-supervised splits within a single dataset (ACDC, Pancreas-NIH, LA), with labeled and unlabeled volumes drawn from the same underlying distribution. No cross-site, cross-scanner, or cross-modality experiment is presented. The transferability claim is therefore not tested; either add such experiments or restrict the claims to the within-dataset semi-supervised setting.
  5. [Section 2.2, Eqs. (2)–(4)] The load-bearing premise that the teacher's class-conditional covariance Σ_t, estimated by EMA over pseudo-labeled unlabeled features, accurately models the true target distribution is not examined. With only 3–12 labeled volumes in the experiments, pseudo-labels are inevitably noisy, and the paper reports no confidence thresholding, pseudo-label filtering, or robustness analysis. Since the perturbation covariance enters L_tsa directly, the authors should either justify the reliability of these statistics or provide a sensitivity analysis with respect to pseudo-label noise or confidence thresholds.
minor comments (8)
  1. [Section 2.2, Eq. (6)] The bias term b_c is described as an element of R^d, but it is a scalar bias; the notation should be corrected.
  2. [Section 2.2, Eq. (8)] The symbol L_∞ is introduced without definition, and the relationship between L_∞ and the L_tsa used in Eq. (9) is not made explicit.
  3. [Section 2.2] There are several typos and unclear phrases, including 'navie' instead of 'naive', 'different∆µ' with improper spacing, and 'with its labeled preserved' in Eq. (4), which should be rewritten for clarity.
  4. [Figure 3(c–d)] The sensitivity-analysis plots are hard to read: the x-axis for β is not labeled, the curves appear as unreadable rendered blobs, and the caption contains font-encoding glyphs such as '/uni0000001b'; the figure should be regenerated with proper axis labels and legends.
  5. [Section 3.1] The text refers to 'Appendix B' for the training procedure, but no appendix is included in the submitted version; the training details are therefore incomplete.
  6. [Sections 1 and 4] Phrases such as 'establishing a new direction' and 'consistently outperforms state-of-the-art' overstate the incremental empirical results and should be moderated.
  7. [Section 3.1 and Eq. (9)] The claim that L_tsa incurs negligible computational overhead is not quantified; the authors should report runtime or memory comparisons against explicit feature augmentation.
  8. [Section 2.2] Since L_tsa is directly adapted from ISDA [19], the paper should include ISDA as a baseline or explicitly clarify the technical differences beyond the change of application setting.

Circularity Check

0 steps flagged · score 2.0 of 10

No load-bearing circularity: the TSA loss is an acknowledged adaptation of ISDA and is benchmarked against external methods; only minor self-citation to GraphCL remains.

full rationale

TransMedSeg's central derivation (TSA upper bound, Eqs. 7-8) is not circular: the loss is constructed by taking the expected cross-entropy over augmented features and bounding it with Jensen's inequality and linearity of expectation, following ISDA [19], which the paper explicitly cites. The subsequent claim of consistent state-of-the-art performance is benchmarked against independent methods (BCP, SS-Net, UA-MT, etc.) as well as against GraphCL, so the empirical result is externally falsifiable rather than forced by the input. The only self-referential element is that the paper builds on and benchmarks against GraphCL [17], the authors' own prior arXiv work; this is a normal incremental-extension setup and does not make the reported gains tautological, especially since the ablation in Table 4 shows that adding L_tsa changes the scores. The beta sensitivity analysis chooses beta=0.4 on the ACDC data and then reports ACDC numbers, but this is ordinary hyperparameter selection, not a fitted parameter renamed as a prediction. I also note for the record that the theoretical derivation has internal problems: the printed Delta_w_c definition cannot produce the cross-entropy in Eq. 6, the Gaussian expectation in Eq. 8 is never evaluated, and Delta_mu has opposite signs in Section 2.1 and Eq. 3. These are correctness/ambiguity defects, not circularity, because they do not make the derivation equivalent to its own inputs. Overall, there is no significant circularity; the score of 2 reflects only the minor, non-load-bearing self-citation to GraphCL.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

No new physical or architectural entities beyond the 'lightweight memory module', which is a statistics buffer rather than an invented entity with independent evidence. The main assumptions are about the validity of Gaussian perturbation and the reliability of pseudo-label statistics.

free parameters (3)
  • beta = 0.4
    Weight of L_tsa in final loss; chosen on ACDC validation by sensitivity analysis; all metrics best at 0.4.
  • alpha = not specified
    Adaptation strength controlling the scale of the Gaussian perturbation; appears in Eqs 3-4 but no value is reported in experiments.
  • Teacher EMA decay lambda = not specified
    Exponential moving average update for teacher parameters and statistics; no value reported.
assumptions (4)
  • domain assumption Class-conditional feature statistics (mu_s, Sigma_s, mu_t, Sigma_t) live in a comparable feature space across student and teacher.
    The method subtracts and adds these statistics across networks; no alignment or calibration is described (Section 2.2).
  • ad hoc to paper The Gaussian perturbation delta ~ N(alpha Delta_mu, alpha Sigma_t) is a valid model of anatomical domain shift.
    No justification beyond the ISDA analogy; the sign of Delta_mu is inconsistent between Eq 1 and Eq 3.
  • domain assumption Teacher pseudo-labels are accurate enough for EMA feature statistics.
    With 3-10 labeled volumes, pseudo-labels on unlabeled data are noisy; the paper does not filter by confidence or analyze error propagation (Section 3.1).
  • standard math Jensen's inequality and the Strong Law of Large Numbers justify the upper-bound surrogate.
    Standard results; the paper invokes them in Section 2.2 but does not show the closed-form MGF that makes the bound computable.

how reviews work

0 comments
Cite this review

Pith. "Pith review of TransMedSeg: A Transferable Semantic Framework for Semi-Supervised Medical Image Segmentation." pith.science (2026). https://pith.science/paper/GFZPXD7J

@misc{pith2026250514753,
  author       = {Pith},
  title        = {Pith review of: TransMedSeg: A Transferable Semantic Framework for Semi-Supervised Medical Image Segmentation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/GFZPXD7J}},
  note         = {Machine review of arXiv:2505.14753}
}
read the original abstract

Semi-supervised learning (SSL) has achieved significant progress in medical image segmentation (SSMIS) through effective utilization of limited labeled data. While current SSL methods for medical images predominantly rely on consistency regularization and pseudo-labeling, they often overlook transferable semantic relationships across different clinical domains and imaging modalities. To address this, we propose TransMedSeg, a novel transferable semantic framework for semi-supervised medical image segmentation. Our approach introduces a Transferable Semantic Augmentation (TSA) module, which implicitly enhances feature representations by aligning domain-invariant semantics through cross-domain distribution matching and intra-domain structural preservation. Specifically, TransMedSeg constructs a unified feature space where teacher network features are adaptively augmented towards student network semantics via a lightweight memory module, enabling implicit semantic transformation without explicit data generation. Interestingly, this augmentation is implicitly realized through an expected transferable cross-entropy loss computed over the augmented teacher distribution. An upper bound of the expected loss is theoretically derived and minimized during training, incurring negligible computational overhead. Extensive experiments on medical image datasets demonstrate that TransMedSeg outperforms existing semi-supervised methods, establishing a new direction for transferable representation learning in medical image analysis.

Figures

Figures reproduced from arXiv: 2505.14753 by the authors.

Figure 1
Figure 1. Overview of TransMedSeg. TransMedSeg enhances the semantic alignment of source [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Illustration of TransMedSeg. For each class, we leverage the inter-domain feature mean [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. (a–b) T-SNE visualization of feature embeddings w/o and with [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Qualitative visualization results from the ACDC dataset. [PITH_FULL_IMAGE:figures/full_fig_p009_4.png]
Figure 5
Figure 5. Figure 5: Qualitative visualization results from the Pancreas-NIH dataset. [PITH_FULL_IMAGE:figures/full_fig_p009_5.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

28 extracted references · 26 canonical work pages

  1. [17]

    Graphcl: Graph-based clustering for semi-supervised medical image segmentation.arXiv preprint arXiv:2411.13147, 2024

    Mengzhu Wang, Jiao Li, Houcheng Su, Nan Yin, Liang Yang, and Shen Li. Graphcl: Graph-based clustering for semi-supervised medical image segmentation.arXiv preprint arXiv:2411.13147, 2024

    arXiv 2024

  2. [1]

    Medical image segmentation review: The success of u-net.IEEE Transactions on Pattern Analysis and Machine Intelligence, 2024

    Reza Azad, Ehsan Khodapanah Aghdam, Amelie Rauland, Yiwei Jia, Atlas Haddadi Avval, Afshin Bozorgpour, Sanaz Karimijafarbigloo, Joseph Paul Cohen, Ehsan Adeli, and Dorit Merhof. Medical image segmentation review: The success of u-net.IEEE Transactions on Pattern Analysis and Machine Intelligence, 2024

    work page 2024

  3. [2]

    Bidirectional copy-paste for semi-supervised medical image segmentation

    Yunhao Bai, Duowen Chen, Qingli Li, Wei Shen, and Yan Wang. Bidirectional copy-paste for semi-supervised medical image segmentation. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 11514–11524, 2023

    work page 2023

  4. [3]

    An exceed- ingly simple consistency regularization method for semi-supervised medical image segmentation

    Hritam Basak, Rajarshi Bhattacharya, Rukhshanda Hussain, and Agniv Chatterjee. An exceed- ingly simple consistency regularization method for semi-supervised medical image segmentation. In2022 IEEE 19th International Symposium on Biomedical Imaging (ISBI), pages 1–4, 2022

    work page 2022

  5. [4]

    Semi-supervised medical image segmentation via learning consistency under transfor- mations

    Gerda Bortsova, Florian Dubost, Laurens Hogeweg, Ioannis Katramados, and Marleen De Brui- jne. Semi-supervised medical image segmentation via learning consistency under transfor- mations. InInternational Conference on Medical Image Computing and Computer-assisted Intervention, pages 810–818, 2019

    work page 2019

  6. [5]

    An elementary proof of the strong law of large numbers.Zeitschrift für Wahrscheinlichkeitstheorie und verwandte Gebiete, 55(1):119–122, 1981

    Nasrollah Etemadi. An elementary proof of the strong law of large numbers.Zeitschrift für Wahrscheinlichkeitstheorie und verwandte Gebiete, 55(1):119–122, 1981

    work page 1981

  7. [6]

    Shape-aware semi-supervised 3d semantic seg- mentation for medical images

    Shuailin Li, Chuyu Zhang, and Xuming He. Shape-aware semi-supervised 3d semantic seg- mentation for medical images. InInternational Conference on Medical Image Computing and Computer-assisted Intervention, pages 552–561, 2020

    work page 2020

  8. [7]

    Semi-supervised medical image segmentation through dual-task consistency

    Xiangde Luo, Jieneng Chen, Tao Song, and Guotai Wang. Semi-supervised medical image segmentation through dual-task consistency. InProceedings of the AAAI conference on artificial intelligence, volume 35, pages 8801–8809, 2021

    work page 2021

Show all 28 references
  1. [8]

    Efficient semi-supervised gross target volume of nasopharyngeal carcinoma segmentation via uncertainty rectified pyramid consistency

    Xiangde Luo, Wenjun Liao, Jieneng Chen, Tao Song, Yinan Chen, Shichuan Zhang, Nianyong Chen, Guotai Wang, and Shaoting Zhang. Efficient semi-supervised gross target volume of nasopharyngeal carcinoma segmentation via uncertainty rectified pyramid consistency. In International ...

    2021

  2. [9]

    Semi-supervised medical image segmentation via uncertainty rectified pyramid consistency.Medical Image Analysis, 80:102517, 2022

    Xiangde Luo, Guotai Wang, Wenjun Liao, Jieneng Chen, Tao Song, Yinan Chen, Shichuan Zhang, Dimitris N Metaxas, and Shaoting Zhang. Semi-supervised medical image segmentation via uncertainty rectified pyramid consistency.Medical Image Analysis, 80:102517, 2022

    2022

  3. [10]

    Jensen’s inequality

    Edward James McShane. Jensen’s inequality. 1937

    1937

  4. [11]

    V-net: Fully convolutional neural networks for volumetric medical image segmentation

    Fausto Milletari, Nassir Navab, and Seyed-Ahmad Ahmadi. V-net: Fully convolutional neural networks for volumetric medical image segmentation. InInternational Conference on 3D Vision, pages 565–571, 2016

    2016

  5. [12]

    Pseudo labeling methods for semi-supervised semantic segmentation: A review and future perspectives.IEEE Transactions on Circuits and Systems for Video Technology, 2024

    Lingyan Ran, Yali Li, Guoqiang Liang, and Yanning Zhang. Pseudo labeling methods for semi-supervised semantic segmentation: A review and future perspectives.IEEE Transactions on Circuits and Systems for Video Technology, 2024

    2024

  6. [13]

    U-net: Convolutional networks for biomedical image segmentation

    Olaf Ronneberger, Philipp Fischer, and Thomas Brox. U-net: Convolutional networks for biomedical image segmentation. InInternational Conference on Medical Image Computing and Computer-assisted Intervention, pages 234–241, 2015

    2015

  7. [14]

    Inconsistency-aware uncertainty estimation for semi-supervised medical image segmentation

    Yinghuan Shi, Jian Zhang, Tong Ling, Jiwen Lu, Yefeng Zheng, Qian Yu, Lei Qi, and Yang Gao. Inconsistency-aware uncertainty estimation for semi-supervised medical image segmentation. IEEE Transactions on Medical Imaging, 41(3):608–620, 2021

    2021

  8. [15]

    Consistency and adversarial semi-supervised learning for medical image segmentation.Computers in Biology and Medicine, 161:107018, 2023

    Yongqiang Tang, Shilei Wang, Yuxun Qu, Zhihua Cui, and Wensheng Zhang. Consistency and adversarial semi-supervised learning for medical image segmentation.Computers in Biology and Medicine, 161:107018, 2023. 10

    2023

  9. [16]

    Advent: Ad- versarial entropy minimization for domain adaptation in semantic segmentation

    Tuan-Hung Vu, Himalaya Jain, Maxime Bucher, Matthieu Cord, and Patrick Pérez. Advent: Ad- versarial entropy minimization for domain adaptation in semantic segmentation. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 2517–2526, 2019

    2019

  10. [18]

    Equity in unsupervised domain adaptation by nuclear norm maximization.IEEE Transactions on Circuits and Systems for Video Technology, 34(7):5533–5545, 2024

    Mengzhu Wang, Shanshan Wang, Xun Yang, Jianlong Yuan, and Wenju Zhang. Equity in unsupervised domain adaptation by nuclear norm maximization.IEEE Transactions on Circuits and Systems for Video Technology, 34(7):5533–5545, 2024

    2024

  11. [19]

    Regularizing deep networks with semantic data augmentation.IEEE Transactions on Pattern Analysis and Machine Intelligence, 44(7):3733–3748, 2021

    Yulin Wang, Gao Huang, Shiji Song, Xuran Pan, Yitong Xia, and Cheng Wu. Regularizing deep networks with semantic data augmentation.IEEE Transactions on Pattern Analysis and Machine Intelligence, 44(7):3733–3748, 2021

    2021

  12. [20]

    Federated semi-supervised medical image segmentation via prototype-based pseudo-labeling and contrastive learning.IEEE Transactions on Medical Imaging, 43(2):649–661, 2023

    Huisi Wu, Baiming Zhang, Cheng Chen, and Jing Qin. Federated semi-supervised medical image segmentation via prototype-based pseudo-labeling and contrastive learning.IEEE Transactions on Medical Imaging, 43(2):649–661, 2023

    2023

  13. [21]

    Semi-supervised left atrium segmentation with mutual consistency training

    Yicheng Wu, Minfeng Xu, Zongyuan Ge, Jianfei Cai, and Lei Zhang. Semi-supervised left atrium segmentation with mutual consistency training. InInternational Conference on Medical Image Computing and Computer-assisted Intervention, pages 297–306, 2021

    2021

  14. [22]

    Mutual consistency learning for semi-supervised medical image segmentation.Medical Image Analysis, 81:102530, 2022

    Yicheng Wu, Zongyuan Ge, Donghao Zhang, Minfeng Xu, Lei Zhang, Yong Xia, and Jianfei Cai. Mutual consistency learning for semi-supervised medical image segmentation.Medical Image Analysis, 81:102530, 2022

    2022

  15. [23]

    Exploring smooth- ness and class-separation for semi-supervised medical image segmentation

    Yicheng Wu, Zhonghua Wu, Qianyi Wu, Zongyuan Ge, and Jianfei Cai. Exploring smooth- ness and class-separation for semi-supervised medical image segmentation. InInternational Conference on Medical Image Computing and Computer-assisted Intervention, pages 34–43, 2022

    2022

  16. [24]

    Segmamba: Long-range sequential modeling mamba for 3d medical image segmentation

    Zhaohu Xing, Tian Ye, Yijun Yang, Guang Liu, and Lei Zhu. Segmamba: Long-range sequential modeling mamba for 3d medical image segmentation. InInternational Conference on Medical Image Computing and Computer-assisted Intervention, pages 578–588, 2024

    2024

  17. [25]

    Pca: Semi-supervised segmentation with patch confidence adversarial training.IEEE Transactions on Network Science and Engineering, 2025

    Zhenghua Xu, Runhe Yang, Zihang Xu, Shuo Zhang, Yuchen Yang, Weipeng Liu, Weichao Xu, Junyang Chen, Thomas Lukasiewicz, and Victor CM Leung. Pca: Semi-supervised segmentation with patch confidence adversarial training.IEEE Transactions on Network Science and Engineering, 2025

    2025

  18. [26]

    Uncertainty- aware self-ensembling model for semi-supervised 3d left atrium segmentation

    Lequan Yu, Shujun Wang, Xiaomeng Li, Chi-Wing Fu, and Pheng-Ann Heng. Uncertainty- aware self-ensembling model for semi-supervised 3d left atrium segmentation. InInternational Conference on Medical Image Computing and Computer-assisted Intervention, pages 605–613, 2019

    2019

  19. [27]

    Deep adversarial networks for biomedical image segmentation utilizing unannotated images

    Yizhe Zhang, Lin Yang, Jianxu Chen, Maridel Fredericksen, David P Hughes, and Danny Z Chen. Deep adversarial networks for biomedical image segmentation utilizing unannotated images. InInternational Conference on Medical Image Computing and Computer-assisted Intervention, pages...

    2017

  20. [28]

    Alternate diverse teaching for semi-supervised medical image segmentation

    Zhen Zhao, Zicheng Wang, Longyue Wang, Dian Yu, Yixuan Yuan, and Luping Zhou. Alternate diverse teaching for semi-supervised medical image segmentation. InEuropean Conference on Computer Vision, pages 227–243, 2024. 11

    2024

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.