Pith. sign in

REVIEW 3 major objections 4 minor 50 references

EAUWSeg: Eliminating annotation uncertainty in weakly-supervised medical image segmentation

T0 review · 3 major / 4 minor · reviewed 2026-08-10 · deepseek-v4-flash

Pith's one-line read This paper claims that labeling a lesion with two rough polygons—one inside, one outside the true boundary—and training with a confidence-guided contrastive consistency learner lets weakly supervised segmentation match or exceed fully…

desk verdict Good idea, broken evidence: the BPAnno masks are GT-derived and the core confidence formulas are internally inconsistent. read the letter →

arxiv 2501.01658 v1 pith:IN5DTGYR submitted 2025-01-03 cs.CV cs.AI

classification cs.CVcs.AI
keywords weakly-supervisedsegmentationboundedpolygonannotationmedicalimagecontrastivelearningconsistencyuncertaintyskinlesionpolyp
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

EAUWSeg sets out to close the gap between weakly supervised and fully supervised medical image segmentation by changing the annotation itself: instead of a dense pixel mask, a specialist draws two rough polygons around a lesion, one inscribed and one envelope-like. Because the pixels between the two polygons are genuinely uncertain, the paper argues that treating the two polygons as separate adversarial labels and then using a classification-guided confidence generator plus a contrastive consistency learner can turn most of that uncertain band into reliable supervision. On ISIC2017 and Kvasir-SEG the resulting models match or beat fully supervised counterparts while requiring roughly 18% of the annotation time, which is why the claim matters: it promises near-full supervision accuracy at a fraction of the labeling cost.

What carries the argument

The load-bearing object is the bounded polygon annotation (BPAnno), which splits the image into certain foreground $\Omega_I$ (inside the inscribed polygon), certain background $\Omega_O$ (outside the envelope polygon), and an uncertain band $\Omega_\Delta$ between them. The framework then uses: (1) two dice losses treating the two polygons as separate labels, so the band provides an adversarial training signal; (2) a classification-guided confidence generator (CCG) that predicts a three-class label per pixel and combines classification confidence with predictive entropy into a confidence map $U$; and (3) a confidence-auxiliary consistency learner (CCL) that runs pixel-wise contrastive learning using only certain pixels and confident uncertain pixels, storing certain-class embeddings in a memory queue. The contrastive loss is the mechanism that actually propagates reliable supervision into the uncertain band; CCG decides which uncertain pixels deserve to be treated as certain.

What would settle it

Measure the CCG's three-class predictions in the $\Omega_\Delta$ band against dense ground-truth masks on a held-out set: if the assigned 'certain' labels in the band agree with ground truth no better than chance, then the confidence-guided contrastive learning is reinforcing noise and the reported gains must come from some other component.

Watch

Extended reading notes

Core claim

The central discovery is that a binary lesion segmentation can be supervised by two nested polygon annotations that bracket the true boundary, and the uncertainty they create can be mostly resolved during training rather than by extra annotation. The paper's EAUWSeg framework trains a segmentation network with two dice losses, one for the inscribed polygon and one for the envelope polygon, so pixels between them receive contradictory labels that push the model to learn boundary-invariant features. A multi-class classification head (background, uncertain band, foreground) then estimates the confidence of each uncertain pixel, and a pixel-wise contrastive loss with a memory queue pulls same-class uncertain pixels toward certain-class features while pushing different classes apart. Reported results on ISIC2017 and Kvasir-SEG show Dice scores of 86.60% and 89.88% with TransUNet, above the fully supervised baselines in the same table, and the modules used to generate confidence are discarded at inference so runtime is unchanged.

Load-bearing premise

The method assumes that the classification head and the entropy filter can reliably identify which pixels in the band between the two polygons are true lesion and which are background, so that the contrastive learner is pulling correct samples together rather than amplifying mislabeled pixels.

Editorial extensions

If this is right

  • Specialists can annotate skin lesions and polyps by drawing two rough polygons instead of a dense mask, cutting annotation time to roughly 18% while keeping or improving Dice scores.
  • Because CCL and CCG are removed at inference, the trained model runs at the same cost as the base segmentation network such as UNet or TransUNet.
  • The bounded-annotation strategy transfers across backbone architectures and, when trained on ISIC2017, generalizes to ISIC2018 without fine-tuning at nearly the same Dice as fully supervised models.
  • On both datasets, EAUWSeg outperforms scribble-, box-, and rectangle-supervised baselines, so the two-polygon annotation is a stronger weak-label format than those alternatives.

Reading between the lines

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

  • As an extension, the two-polygon annotation could be re-used as a general weak-label format: any scribble or bounding-box annotation can be expanded into an inner and outer region, and the same CCG/CCL machinery may convert the uncertain band into supervision.
  • The 18% annotation-time figure comes from a single dermatologist on 20 ISIC2017 images; a multi-rater study with varied lesion shapes and image qualities would show whether the workload saving holds in practice.
  • If the confidence generator is reliable, EAUWSeg should combine naturally with semi-supervised learning on unlabeled images; the remaining gap to the best semi-supervised method suggests that unlabeled data could push it past fully supervised performance.
  • A straightforward testable extension is to apply the same bounded-polygon supervision to multi-class segmentation; the paper itself notes that an envelope polygon containing multiple foreground classes is an unsolved challenge.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 4 minor

Summary. The paper proposes BPAnno, a weakly-supervised annotation scheme in which each lesion is annotated by two polygons, an inscribed-like polygon and an envelope-like polygon, so that the region between them is treated as uncertain. The accompanying EAUWSeg framework explicitly treats the two polygons as separate supervision signals, and adds a classification-guided confidence generator (CCG) and a confidence-auxiliary consistency learner (CCL) to produce pseudo-supervision for the uncertain band. Experiments on ISIC2017 and Kvasir-SEG compare EAUWSeg with scribble-, box-, and fully-supervised baselines, and the paper reports that EAUWSeg outperforms fully supervised training with less than 20% of the annotation workload. The paper also includes ablations, a semi-supervised comparison, a cross-dataset generalization study on ISIC2018, and an annotation-cost study.

Significance. The bounded-polygon annotation idea is potentially attractive: it is a natural middle ground between dense masks and boxes, and the proposed framework is modular and applicable to standard segmentation backbones. The paper provides extensive experiments, ablations across components, backbone generalization, and a cross-dataset generalization check, which are useful strengths. However, the central empirical claim is not supported by the current evaluation protocol: the BPAnno masks used for training are generated from the ground-truth masks, so the comparison against fully supervised training does not test the claimed manual two-polygon annotation workflow. In addition, the confidence-generator equations are internally inconsistent, so the core mechanism is underspecified. If both issues were fixed, the method could be of interest, but as submitted the headline contributions are not established.

major comments (3)
  1. [§IV-A2, Table I, §IV-E] The central empirical claim is undermined by the annotation-generation protocol. Section IV-A2 states that the bounded-polygon annotations are initially generated by dilation and erosion of the available ground-truth masks, then simplified with the Douglas-Peucker algorithm and manually refined. Thus the BPAnno masks used to train the models in Table I are not independent weak annotations drawn by a clinician; they are a ground-truth-informed envelope around the exact masks used by the fully-supervised baselines. The abstract and conclusion claim that EAUWSeg 'outperforms fully-supervised counterparts' with 'less than 20% of the annotation workload,' but the comparison cannot separate the method's contribution from leaked boundary information in the supervision masks. The annotation-cost study in Section IV-E uses one dermatologist on twenty images and never verifies that manually drawn polygons agree with the ground-truth-derived polygons used in the actual training experiments, so it does not repair this gap.
  2. [§III-C, Eqs. (10)-(11)] The classification-guided confidence generator is not internally consistent. Eq. (10) defines U_c = argmax(P(y=0|f_S(x), Θ_c), P(y=2|f_S(x), Θ_c)) ⊙ M_u, which takes the argmax over only the background and foreground classes. Consequently, it cannot ever mark a pixel as belonging to class 1, even though the text says that pixels with prediction of 1 for the multi-class classification task are treated as solid uncertain. Moreover, argmax returns a class index, not a confidence value. The situation is compounded in Eq. (11): U = min(U_c + 2U_e, -1) ⊙ M_u, where U_c ∈ {0,2} and U_e ∈ {0,-1}; with these ranges the expression collapses to -1 or -2 for nearly all combinations, making the intended distinction between 'solid uncertain' and 'certain' pixels ill-defined. Because the confidence-auxiliary consistency learner in Section III-D selects pseudo-labels from this map, the central mechanism is not reproducibly specified as written.
  3. [§III-D, Eq. (12)] The pseudo-label construction inherits the inconsistency of the confidence map. In Eq. (12), for pixels in the uncertain band (M_u=1), the pseudo-label becomes U, which can take value -2 according to Eq. (11); however, the text only describes -1 as the label for solid uncertain pixels. The pixel-wise contrastive loss in Eq. (13) then treats these values as class labels for positive/negative sampling, but no rule is given for interpreting -2 or for deciding whether such pixels should be included, excluded, or reweighted. This makes the sample-selection criterion for the contrastive learner incomplete.
minor comments (4)
  1. [§III-C, Eq. (8)] Eq. (8) declares U_e to be in R^{C×H×W}, but the right-hand side assigns scalar values of -1 and 0; please clarify the intended shape and how the uncertainty map relates to the channel dimension.
  2. [Fig. 2] The figure caption and diagram use terms such as 'solid classes,' 'solid error predictions,' and 'Store the embedding in solid classes' without defining 'solid' in the text; please define the terminology consistently.
  3. [Abstract and Conclusion] The abstract and conclusion state that EAUWSeg 'outperforms fully-supervised counterparts,' but Section IV-C5 and Table V report that the cross-dataset generalization performance is 'comparable' rather than superior; the wording should be aligned with the actual results.
  4. [§IV-A2] The scribble annotations for the comparison are generated by connecting random endpoints sampled from the ground-truth mask, which may not reflect how clinicians draw scribbles in practice; this should be acknowledged as a limitation of the comparison protocol.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity: the reported results are standard test-set evaluations, and the few self-citations are not load-bearing; the GT-derived BPAnno masks raise a validity concern, not a circularity concern.

full rationale

The paper's central derivation is not circular. The bounded-polygon annotation is defined geometrically (two polygons per lesion), and the training losses in Eqs. (4)-(15) are derived from those polygon annotations rather than from the test ground-truth or from the reported evaluation metrics. The Dice, Jaccard, Accuracy, and Sensitivity numbers in Tables I, IV, and V are obtained by evaluating on held-out test sets, so the benchmark results are not constructed to equal the method's inputs. There are two self-citations (refs. 15 and 48), but neither is load-bearing: ref. 15 supports only a general statement that image-level labels give limited performance, and ref. 48 is cited only for the cross-training evaluation protocol. No uniqueness theorem or central premise is imported from the authors' prior work. The largest concern in the paper is experimental rather than circular: Section IV-A2 states that the BPAnno masks were initially generated by dilation-erosion of the available ground-truth masks and then manually refined, so the training masks used in Table I are not produced by the claimed manual annotation workflow. This could leak boundary information and weaken the claim of superiority over fully supervised training, but it is a validity or leakage issue, not a reduction of the claimed result to its inputs by definition. Likewise, the apparent internal inconsistency in Eqs. (10)-(11) for the confidence map is a correctness problem, not a circularity problem. Because no load-bearing step reduces to its own inputs, and because the test-set evaluations are standard, the circularity score is low.

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

The method introduces no new physical or geometric objects. The free parameters are mostly standard training hyperparameters, but the unreported entropy threshold and polygon generation details are load-bearing. The axioms are empirical assumptions about feature-space consistency and the annotation geometry.

free parameters (4)
  • Entropy threshold mu = not reported
    Defines which pixels are solid uncertain in Eq. 8. Never stated in the paper, so the size of the masked region is a free hidden parameter.
  • Loss weights lambda1, lambda2 = 0.3, 0.5
    Weights for contrastive and classification losses in Eq. 15. No sensitivity analysis is provided.
  • Temperature tau = 0.1
    Temperature for pixel contrastive loss in Eq. 13. Set without sensitivity analysis.
  • Polygon generation parameters = not reported
    Dilation/erosion kernel sizes and Douglas-Peucker epsilon used to generate BPAnno from groundtruth masks are not specified, affecting the annotation uncertainty band.
assumptions (3)
  • domain assumption Similar pixels in feature space tend to produce consistent category predictions.
    Invoked in Section III-C via reference [10] to justify using feature similarity to label uncertain pixels.
  • domain assumption The two polygons bracket the lesion boundary, so the uncertain band between them is the only region of label ambiguity.
    Assumed throughout the method; if the envelope polygon misses part of the lesion or the inscribed polygon includes background, the certain-region supervision would be corrupted.
  • ad hoc to paper The classification head can learn meaningful foreground/background/uncertain categories from the bounded polygons alone.
    The classification-guided confidence generator relies on this assumption; no independent evidence is given that the learned categories transfer to the uncertain band.

how reviews work

0 comments
Cite this review

Pith. "Pith review of EAUWSeg: Eliminating annotation uncertainty in weakly-supervised medical image segmentation." pith.science (2026). https://pith.science/paper/IN5DTGYR

@misc{pith2026250101658,
  author       = {Pith},
  title        = {Pith review of: EAUWSeg: Eliminating annotation uncertainty in weakly-supervised medical image segmentation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/IN5DTGYR}},
  note         = {Machine review of arXiv:2501.01658}
}
read the original abstract

Weakly-supervised medical image segmentation is gaining traction as it requires only rough annotations rather than accurate pixel-to-pixel labels, thereby reducing the workload for specialists. Although some progress has been made, there is still a considerable performance gap between the label-efficient methods and fully-supervised one, which can be attributed to the uncertainty nature of these weak labels. To address this issue, we propose a novel weak annotation method coupled with its learning framework EAUWSeg to eliminate the annotation uncertainty. Specifically, we first propose the Bounded Polygon Annotation (BPAnno) by simply labeling two polygons for a lesion. Then, the tailored learning mechanism that explicitly treat bounded polygons as two separated annotations is proposed to learn invariant feature by providing adversarial supervision signal for model training. Subsequently, a confidence-auxiliary consistency learner incorporates with a classification-guided confidence generator is designed to provide reliable supervision signal for pixels in uncertain region by leveraging the feature presentation consistency across pixels within the same category as well as class-specific information encapsulated in bounded polygons annotation. Experimental results demonstrate that EAUWSeg outperforms existing weakly-supervised segmentation methods. Furthermore, compared to fully-supervised counterparts, the proposed method not only delivers superior performance but also costs much less annotation workload. This underscores the superiority and effectiveness of our approach.

Figures

Figures reproduced from arXiv: 2501.01658 by the authors.

Figure 1
Figure 1. Comparison of the typical weak annotation methods and our proposed [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. The overall framework of our proposed EAUWSeg. It includes a segmentation model supervised by two bounded polygons and a multi-class [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Qualitative comparison of different methods on ISIC2017 (top three rows) and Kvasir-SEG (bottom three rows). The green and blue contours indicate [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Performance comparison of EAUWSeg combined with different [PITH_FULL_IMAGE:figures/full_fig_p008_4.png]
Figure 6
Figure 6. Figure 6: t-SNE visualizations on ISIC2017 test set for the TransUNet and [PITH_FULL_IMAGE:figures/full_fig_p009_6.png]
Figure 5
Figure 5. Figure 5: Error analysis on the ISIC2017 test set. Both inside and outside a [PITH_FULL_IMAGE:figures/full_fig_p009_5.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

50 extracted references · 40 canonical work pages

  1. [1]

    Dmsps: Dynamically mixed soft pseudo-label supervision for scribble-supervised medical image segmentation,

    M. Han, X. Luo, X. Xie, W. Liao, S. Zhang, T. Song, G. Wang, and S. Zhang, “Dmsps: Dynamically mixed soft pseudo-label supervision for scribble-supervised medical image segmentation,” Medical Image Analysis, p. 103274, 2024. 10

  2. [2]

    Pa-seg: Learning from point annotations for 3d medical image segmentation using contextual regularization and cross knowledge distillation,

    S. Zhai, G. Wang, X. Luo, Q. Yue, K. Li, and S. Zhang, “Pa-seg: Learning from point annotations for 3d medical image segmentation using contextual regularization and cross knowledge distillation,” IEEE Transactions on Medical Imaging , 2023

  3. [3]

    Segmentation only uses sparse annotations: Unified weakly and semi-supervised learning in medical images,

    F. Gao, M. Hu, M.-E. Zhong, S. Feng, X. Tian, X. Meng, Z. Huang, M. Lv, T. Song, X. Zhang, X. Zou, and X. Wu, “Segmentation only uses sparse annotations: Unified weakly and semi-supervised learning in medical images,” Medical Image Analysis , vol. 80, p. 102515, 2022

  4. [4]

    Weakly supervised brain lesion segmentation via attentional representation learn- ing,

    K. Wu, B. Du, M. Luo, H. Wen, Y . Shen, and J. Feng, “Weakly supervised brain lesion segmentation via attentional representation learn- ing,” in Medical Image Computing and Computer Assisted Intervention . Springer, 2019, pp. 211–219

  5. [5]

    Scribblesup: Scribble- supervised convolutional networks for semantic segmentation,

    D. Lin, J. Dai, J. Jia, K. He, and J. Sun, “Scribblesup: Scribble- supervised convolutional networks for semantic segmentation,” in Pro- ceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2016, pp. 3159–3167

  6. [6]

    Deepcut: Object segmentation from bounding box annotations using convolutional neural networks,

    M. Rajchl, M. C. Lee, O. Oktay, K. Kamnitsas, J. Passerat-Palmbach, W. Bai, M. Damodaram, M. A. Rutherford, J. V . Hajnal, B. Kainz et al., “Deepcut: Object segmentation from bounding box annotations using convolutional neural networks,” IEEE Transactions on Medical Imaging, vol. 36, no. 2, pp. 674–683, 2016

  7. [7]

    Inter extreme points geodesics for end-to-end weakly supervised image segmentation,

    R. Dorent, S. Joutard, J. Shapey, A. Kujawa, M. Modat, S. Ourselin, and T. Vercauteren, “Inter extreme points geodesics for end-to-end weakly supervised image segmentation,” in Medical Image Computing and Computer Assisted Intervention . Springer, 2021, pp. 615–624

  8. [8]

    Scribblevc: Scribble- supervised medical image segmentation with vision-class embedding,

    Z. Li, Y . Zheng, X. Luo, D. Shan, and Q. Hong, “Scribblevc: Scribble- supervised medical image segmentation with vision-class embedding,” in Proceedings of the 31st ACM International Conference on Multimedia , 2023, pp. 3384–3393

Show all 50 references
  1. [9]

    Blpseg: Balance the label preference in scribble-supervised semantic segmentation,

    Y . Wang, J. Zhang, M. Kan, S. Shan, and X. Chen, “Blpseg: Balance the label preference in scribble-supervised semantic segmentation,” IEEE Transactions on Image Processing , 2023

  2. [10]

    Sparsely annotated semantic segmentation with adaptive gaussian mixtures,

    L. Wu, Z. Zhong, L. Fang, X. He, Q. Liu, J. Ma, and H. Chen, “Sparsely annotated semantic segmentation with adaptive gaussian mixtures,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2023, pp. 15 454–15 464

  3. [11]

    N. C. Codella, D. Gutman, M. E. Celebi, B. Helba, M. A. Marchetti, S. W. Dusza, A. Kalloo, K. Liopyris, N. Mishra, H. Kittler, and A. Halpern, “Skin lesion analysis toward melanoma detection: A chal- lenge at the 2017 international symposium on biomedical imaging (isbi), hoste...

  4. [12]

    Kvasir-seg: A segmented polyp dataset,

    D. Jha, P. H. Smedsrud, M. A. Riegler, P. Halvorsen, T. de Lange, D. Johansen, and H. D. Johansen, “Kvasir-seg: A segmented polyp dataset,” in MultiMedia Modeling: 26th International Conference, MMM

  5. [13]

    Embracing imperfect datasets: A review of deep learning solutions for medical image segmentation,

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

  6. [14]

    Learning deep features for discriminative localization,

    B. Zhou, A. Khosla, A. Lapedriza, A. Oliva, and A. Torralba, “Learning deep features for discriminative localization,” inProceedings of the IEEE Conference on Computer Vision and Pattern Recognition , 2016, pp. 2921–2929

  7. [15]

    Intra-class consistency and inter-class discrimination feature learning for automatic skin lesion classification,

    L. Wang, L. Zhang, X. Shu, and Z. Yi, “Intra-class consistency and inter-class discrimination feature learning for automatic skin lesion classification,” Medical Image Analysis , vol. 85, p. 102746, 2023

  8. [16]

    Weakly supervised instance segmentation using the bounding box tightness prior,

    C.-C. Hsu, K.-J. Hsu, C.-C. Tsai, Y .-Y . Lin, and Y .-Y . Chuang, “Weakly supervised instance segmentation using the bounding box tightness prior,” Advances in Neural Information Processing Systems , vol. 32, 2019

  9. [17]

    Learning to segment from scribbles using multi-scale adversarial attention gates,

    G. Valvano, A. Leo, and S. A. Tsaftaris, “Learning to segment from scribbles using multi-scale adversarial attention gates,” IEEE Transac- tions on Medical Imaging , vol. 40, no. 8, pp. 1990–2001, 2021

  10. [18]

    Scribble-based hierarchical weakly supervised learning for brain tumor segmentation,

    Z. Ji, Y . Shen, C. Ma, and M. Gao, “Scribble-based hierarchical weakly supervised learning for brain tumor segmentation,” in Medical Image Computing and Computer Assisted Intervention . Springer, 2019, pp. 175–183

  11. [19]

    Cyclemix: A holistic strategy for medical image segmentation from scribble supervision,

    K. Zhang and X. Zhuang, “Cyclemix: A holistic strategy for medical image segmentation from scribble supervision,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , 2022, pp. 11 656–11 665

  12. [20]

    Simclr: A simple framework for contrastive learning of visual representations,

    T. Chen, S. Kornblith, M. Norouzi, and G. Hinton, “Simclr: A simple framework for contrastive learning of visual representations,” in Inter- national Conference on Learning Representations , vol. 2, 2020

  13. [21]

    A survey on contrastive self-supervised learning,

    A. Jaiswal, A. R. Babu, M. Z. Zadeh, D. Banerjee, and F. Makedon, “A survey on contrastive self-supervised learning,” Technologies, vol. 9, no. 1, p. 2, 2020

  14. [22]

    Rethinking semi-supervised medical image segmentation: A variance-reduction perspective,

    C. You, W. Dai, Y . Min, F. Liu, D. Clifton, S. K. Zhou, L. Staib, and J. Duncan, “Rethinking semi-supervised medical image segmentation: A variance-reduction perspective,” in Advances in Neural Information Processing Systems, vol. 36, 2023, pp. 9984–10 021

  15. [23]

    Exploring cross-image pixel contrast for semantic segmentation,

    W. Wang, T. Zhou, F. Yu, J. Dai, E. Konukoglu, and L. Van Gool, “Exploring cross-image pixel contrast for semantic segmentation,” in Proceedings of the IEEE/CVF International Conference on Computer Vision, 2021, pp. 7303–7313

  16. [24]

    Uncertainty-guided pixel contrastive learning for semi-supervised medical image segmentation,

    T. Wang, J. Lu, Z. Lai, J. Wen, and H. Kong, “Uncertainty-guided pixel contrastive learning for semi-supervised medical image segmentation,” in Proceedings of the Thirty-First International Joint Conference on Artificial Intelligence, 2022, pp. 1444–1450

  17. [25]

    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 Computing and Computer-Assisted Intervention. Springer, 2015, pp. 234–241

  18. [26]

    Encoder- decoder with atrous separable convolution for semantic image segmen- tation,

    L.-C. Chen, Y . Zhu, G. Papandreou, F. Schroff, and H. Adam, “Encoder- decoder with atrous separable convolution for semantic image segmen- tation,” in Proceedings of the European Conference on Computer Vision, 2018, pp. 801–818

  19. [27]

    Transunet: Rethinking the u-net architecture design for medical image segmentation through the lens of transformers,

    J. Chen, J. Mei, X. Li, Y . Lu, Q. Yu, Q. Wei, X. Luo, Y . Xie, E. Adeli, Y . Wang, M. P. Lungren, S. Zhang, L. Xing, L. Lu, A. Yuille, and Y . Zhou, “Transunet: Rethinking the u-net architecture design for medical image segmentation through the lens of transformers,” Medical ...

  20. [28]

    Exploring feature representation learning for semi-supervised medical image segmentation,

    H. Wu, X. Li, and K.-T. Cheng, “Exploring feature representation learning for semi-supervised medical image segmentation,” IEEE Trans- actions on Neural Networks and Learning Systems , 2023

  21. [29]

    Contrastive registra- tion for unsupervised medical image segmentation,

    L. Liu, A. I. Aviles-Rivero, and C.-B. Sch ¨onlieb, “Contrastive registra- tion for unsupervised medical image segmentation,” IEEE Transactions on Neural Networks and Learning Systems , 2023

  22. [30]

    Unet++: Redesigning skip connections to exploit multiscale features in image segmentation,

    Z. Zhou, M. M. R. Siddiquee, N. Tajbakhsh, and J. Liang, “Unet++: Redesigning skip connections to exploit multiscale features in image segmentation,” IEEE Transactions on Medical Imaging , vol. 39, no. 6, pp. 1856–1867, 2019

  23. [31]

    Transfuse: Fusing transformers and cnns for medical image segmentation,

    Y . Zhang, H. Liu, and Q. Hu, “Transfuse: Fusing transformers and cnns for medical image segmentation,” in Medical Image Computing and Computer Assisted Intervention . Springer, 2021, pp. 14–24

  24. [32]

    Hiformer: Hierarchical multi-scale representations using transformers for medical image segmentation,

    M. Heidari, A. Kazerouni, M. Soltany, R. Azad, E. K. Aghdam, J. Cohen-Adad, and D. Merhof, “Hiformer: Hierarchical multi-scale representations using transformers for medical image segmentation,” in Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Visio...

  25. [33]

    Normal- ized cut loss for weakly-supervised cnn segmentation,

    M. Tang, A. Djelouah, F. Perazzi, Y . Boykov, and C. Schroers, “Normal- ized cut loss for weakly-supervised cnn segmentation,” in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition , 2018, pp. 1818–1827

  26. [34]

    Unsupervised total variation loss for semi-supervised deep learning of semantic segmenta- tion,

    M. Javanmardi, M. Sajjadi, T. Liu, and T. Tasdizen, “Unsupervised total variation loss for semi-supervised deep learning of semantic segmenta- tion,” arXiv preprint arXiv:1605.01368 , 2016

  27. [35]

    Gated crf loss for weakly supervised semantic image segmentation,

    A. Obukhov, S. Georgoulis, D. Dai, and L. Van Gool, “Gated crf loss for weakly supervised semantic image segmentation,” arXiv preprint arXiv:1906.04651, 2019

  28. [36]

    Mumford–shah loss functional for image seg- mentation with deep learning,

    B. Kim and J. C. Ye, “Mumford–shah loss functional for image seg- mentation with deep learning,” IEEE Transactions on Image Processing, vol. 29, pp. 1856–1866, 2019

  29. [37]

    Weakly supervised segmentation of covid19 infection with scribble annotation on ct images,

    X. Liu, Q. Yuan, Y . Gao, K. He, S. Wang, X. Tang, J. Tang, and D. Shen, “Weakly supervised segmentation of covid19 infection with scribble annotation on ct images,”Pattern Recognition, vol. 122, p. 108341, 2022

  30. [38]

    Trimix: A general framework for medical image segmentation from limited supervision,

    Z. Zheng, Y . Hayashi, M. Oda, T. Kitasaka, and K. Mori, “Trimix: A general framework for medical image segmentation from limited supervision,” in Proceedings of the Asian Conference on Computer Vision, 2022, pp. 634–651

  31. [39]

    Skin lesion analysis toward melanoma detection 2018: A challenge hosted by the international skin imaging collaboration (isic),

    N. Codella, V . Rotemberg, P. Tschandl, M. E. Celebi, S. Dusza, D. Gut- man, B. Helba, A. Kalloo, K. Liopyris, M. Marchetti, H. Kittler, and A. Halpern, “Skin lesion analysis toward melanoma detection 2018: A challenge hosted by the international skin imaging collaboration (is...

  32. [40]

    Algorithms for the reduction of the number of points required to represent a digitized line or its caricature,

    D. H. Douglas and T. K. Peucker, “Algorithms for the reduction of the number of points required to represent a digitized line or its caricature,” Cartographica: the International Journal for Geographic Information and Geovisualization, vol. 10, no. 2, pp. 112–122, 1973

  33. [41]

    Scribbleprompt: Fast and flexible interactive segmentation for any medical image,

    H. E. Wong, M. Rakic, J. Guttag, and A. V . Dalca, “Scribbleprompt: Fast and flexible interactive segmentation for any medical image,” arXiv preprint arXiv:2312.07381, 2023. 11

  34. [42]

    Deep residual learning for image recognition,

    K. He, X. Zhang, S. Ren, and J. Sun, “Deep residual learning for image recognition,” in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2016, pp. 770–778

  35. [43]

    Consistency and adversarial semi-supervised learning for medical image segmentation,

    Y . Tang, S. Wang, Y . Qu, Z. Cui, and W. Zhang, “Consistency and adversarial semi-supervised learning for medical image segmentation,” Computers in Biology and Medicine , vol. 161, p. 107018, 2023

  36. [44]

    Cross-level contrastive learning and consistency constraint for semi-supervised med- ical image segmentation,

    X. Zhao, C. Fang, D.-J. Fan, X. Lin, F. Gao, and G. Li, “Cross-level contrastive learning and consistency constraint for semi-supervised med- ical image segmentation,” in 2022 IEEE 19th International Symposium on Biomedical Imaging , 2022, pp. 1–5

  37. [45]

    Mean teachers are better role mod- els: Weight-averaged consistency targets improve semi-supervised deep learning results,

    A. Tarvainen and H. Valpola, “Mean teachers are better role mod- els: Weight-averaged consistency targets improve semi-supervised deep learning results,” Advances in Neural Information Processing Systems , vol. 30, 2017

  38. [46]

    St++: Make self- training work better for semi-supervised semantic segmentation,

    L. Yang, W. Zhuo, L. Qi, Y . Shi, and Y . Gao, “St++: Make self- training work better for semi-supervised semantic segmentation,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2022, pp. 4268–4277

  39. [47]

    Semi- supervised semantic segmentation with pixel-level contrastive learning from a class-wise memory bank,

    I. Alonso, A. Sabater, D. Ferstl, L. Montesano, and A. C. Murillo, “Semi- supervised semantic segmentation with pixel-level contrastive learning from a class-wise memory bank,” in Proceedings of the IEEE/CVF International Conference on Computer Vision , 2021, pp. 8219–8228

  40. [48]

    Multi-level attention network for retinal vessel segmentation,

    Y . Yuan, L. Zhang, L. Wang, and H. Huang, “Multi-level attention network for retinal vessel segmentation,” IEEE Journal of Biomedical and Health Informatics , vol. 26, no. 1, pp. 312–323, 2021

  41. [49]

    Boxsup: Exploiting bounding boxes to super- vise convolutional networks for semantic segmentation,

    J. Dai, K. He, and J. Sun, “Boxsup: Exploiting bounding boxes to super- vise convolutional networks for semantic segmentation,” in Proceedings of the IEEE International Conference on Computer Vision , 2015, pp. 1635–1643

  42. [2020]

    Springer, 2020, pp. 451–462

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.