Robust Noisy Pseudo-label Learning for Semi-supervised Medical Image Segmentation Using Diffusion Model

Aldo Rinaldi; Cheng Wang; Kawal S. Rhode; Lin Xi; Sandra Howell; YingLiang Ma

REVIEW 5 major objections 5 minor 35 references

Robust Noisy Pseudo-label Learning for Semi-supervised Medical Image Segmentation Using Diffusion Model

T0 review · 5 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 A prototype-contrastive diffusion decoder beats prior semi-supervised segmentation methods by +5.15 mIoU on Endoscapes2023.

desk verdict Useful new benchmark and a plausible diffusion-segmentation combination, but the SoTA claim needs same-backbone baselines and error bars; worth a referee. read the letter →

arxiv 2507.16429 v1 pith:NCV5EXCU submitted 2025-07-22 cs.CV

Lin Xi , Yingliang Ma , Cheng Wang , Sandra Howell , Aldo Rinaldi , Kawal S. Rhode This is my paper

classification cs.CV

keywords semi-supervisedlearningmedicalimagesegmentationdiffusionmodelspseudo-labelnoiseprototypecontrastiveX-rayangiographyEndoscapes2023MOSXAV

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

This paper tries to make semi-supervised medical image segmentation robust to noisy pseudo-labels by restructuring the latent space of a diffusion-based segmenter. The core idea is to treat segmentation as a label-denoising problem, then anchor learned pixel embeddings to fixed class prototypes with two contrastive losses, one pulling same-class pixels together and one pushing different classes apart. The authors report that this beats state-of-the-art baselines, including MedSegDiff by +5.15 mIoU on Endoscapes2023, and introduce MOSXAV, a new multi-object X-ray angiography segmentation benchmark. If the result holds, noisy pseudo-labels from automatic tools become less damaging, and data-efficient clinical segmentation improves.

What carries the argument

The central object is the diffusion decoder that predicts dense labels by denoising corrupted label embeddings: a label encoder maps a one-hot mask to a normalized embedding z, Gaussian noise with a cosine schedule corrupts it to z_t, and the decoder, conditioned on features from a ViT backbone and visual-specific branches, reconstructs the clean embedding. The second load-bearing piece is the set of fixed, non-learnable class prototypes {p_{c,k}}, updated by online clustering, which serve as anchors in the projected feature space Z = φ(z). The contrastive losses L_inter and L_intra pull each pixel toward its assigned prototype and away from other classes' prototypes, which is what protects dense predictions from the misleading gradients of noisy pseudo-labels. Inference uses DDIM to iterate from noise to a predicted mask.

What would settle it

Retrain all Table 1 baselines with the same ViT-Large backbone, 512x512 input resolution, 40,000 iterations, batch size 8, and the same MedSAM plus SAMRefiner pseudo-label protocol. If the mIoU gap over MedSegDiff shrinks to near zero while the prototype losses are removed from the proposed model, the core claim fails.

Watch

Extended reading notes

Core claim

The paper claims that a conditional diffusion decoder, guided by multi-level visual features, can learn semantic label distributions even when supervision mixes ground truth and noisy pseudo-labels, provided the latent embedding space is explicitly structured. The structuring device is a set of non-learnable class prototypes that act as fixed anchors: pixel embeddings are assigned to prototypes by online clustering, and a prototype-anchored contrastive loss enforces intra-class compactness and inter-class separability. With this mechanism, the model reaches 56.68 mIoU on Endoscapes2023 with GT plus pseudo labels, 55.33 with GT only, and 43.04 on the MOSXAV test set, surpassing MedSegDiff on every setting. The same paper releases MOSXAV, a public benchmark with 62 X-ray angiography video sequences and manually annotated ground truth.

Load-bearing premise

The stated gains assume that the baseline methods were trained under settings as favorable as the proposed method's; if the baselines used smaller backbones, lower input resolution, or fewer iterations, the +5.15 mIoU gap could reflect training budget rather than the prototype-contrastive mechanism.

Editorial extensions

If this is right

Semi-supervised medical segmentation can be improved by anchoring latent semantics to fixed prototypes instead of relying only on mask-level consistency. If the reported gains are reproducible, noisy pseudo-labels become far less damaging during training. MOSXAV gives the community a new public benchmark for multi-object X-ray angiography segmentation, with 2,335 densely annotated training and val

Reading between the lines

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

An implication the paper leaves implicit is that the same prototype-anchored contrastive constraint could be grafted onto non-diffusion segmentation heads, offering a cheaper way to test whether the mechanism, rather than the diffusion decoder, is responsible for the gains. A natural next step is to measure inter-observer agreement on MOSXAV's manual annotations, since the paper does not report it

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.

Desk Editor's Note

read the letter

The one thing to know: this is a solid but uneven paper. The genuinely new piece is MOSXAV, a public benchmark for multi-object X-ray angiography segmentation, and the method reasonably combines diffusion label denoising with prototype contrastive losses. The main empirical claim, however, is not yet controlled enough to trust.

What's actually new: the MOSXAV benchmark is a real contribution—62 videos with manual annotations of vascular structures, publicly linked. That alone merits attention. The technical combination is a sensible extension of DDP/MedSegDiff: anchor the latent space of label embeddings with non-learnable prototypes while training the diffusion decoder. The ablation shows each loss component adds a small, consistent gain, which is honest.

Where it's soft: first, the key comparison is uncontrolled. The proposed method uses ViT-Large; Table 1 lists no backbones for any baseline, so the +5.15 mIoU margin over MedSegDiff could be a capacity effect. There are no error bars, so we can't judge stability. Second, the paper never compares against the semi-supervised methods it cites, such as FixMatch or Cross Pseudo Supervision, even though the setting is explicitly semi-supervised. Third, the MOSXAV description is internally inconsistent: 'dense annotations every 5 frames across 50 sequences' should give roughly 500-700 frames, not 2,335. Fourth, the prototype contrastive loss is applied to the clean label embedding before noise is added, not during the denoising; the abstract's phrasing overstates the connection. Finally, the title promises robustness to noisy pseudo-labels, but there is no experiment that actually perturbs pseudo-label noise.

None of these are fatal. Circularity isn't a concern—the evaluation uses external benchmarks, and the self-citations aren't load-bearing. The method is plausible and the benchmark is useful. But the state-of-the-art claim needs same-backbone baselines, error bars, a direct pseudo-noise robustness test, and a corrected dataset description before it's convincing.

Who this is for: researchers working on semi-supervised medical segmentation, particularly with diffusion models, and anyone needing a new angiography benchmark. It deserves a serious referee, because the benchmark and the idea are worth engaging with, even if the current evidence is conditional.

Recommendation: send it to peer review, with a request for the missing controlled experiments and dataset documentation.

Referee Report

5 major / 5 minor

Summary. The paper proposes a diffusion-based framework for semi-supervised medical image segmentation, combining a conditional diffusion decoder with prototype-anchored contrastive losses on label embeddings. It also introduces MOSXAV, a new multi-object X-ray angiography segmentation benchmark. The method is evaluated on Endoscapes2023 and MOSXAV, and the authors report state-of-the-art results, including a +5.15 mIoU gain over MedSegDiff under the GT+Pseudo setting on Endoscapes2023.

Significance. If the results hold, the contribution is potentially valuable: the paper releases a new public benchmark, applies dense label denoising with structured latent anchors, and provides algorithm pseudocode and a repository link. The dataset release and the external-benchmark evaluation are strengths. However, the current evidence for the state-of-the-art claim is not conclusive. The baseline comparison is uncontrolled with respect to backbone capacity, and the prototype-contrastive losses are applied to the clean ground-truth label embedding rather than to the denoised prediction, which contradicts the claimed mechanism. These issues are fixable but require additional experiments or a substantially revised description.

major comments (5)

[Section 2.3, Algorithm 1] The prototype-contrastive losses are not applied to the denoising process. In Algorithm 1, z_latent = latent_encoder(label_encoder(gts)) is computed from the ground-truth label map before noise is added, and the inter/intra losses in Eq. (2)-(3) are evaluated on this clean embedding; z_crpt, the timestep t, and the predicted preds do not enter these losses. The abstract and Section 2.3 claim a constraint 'during the denoising diffusion process', which does not describe this implementation. Please either move the prototype losses to the predicted/denoised embedding and re-run the experiments, or clearly state that the regularizer is applied to the target label embedding and justify why it should improve denoised predictions.
[Section 3.1, Table 1] The baseline comparison in Table 1 is not controlled for backbone capacity or training protocol. Implementation Details report ViT-Large for the proposed method and state that all models are trained for 40,000 iterations with batch size 8, but no backbone, input resolution, or pseudo-label protocol is reported for U-Net, Attention U-Net, TransU-Net, CMU-Net, CMU-NeXt, SwinU-Net, or MedSegDiff. The +5.15 mIoU margin over MedSegDiff on Endoscapes2023 and the +3.11 mIoU on the MOSXAV test set may therefore stem from the ViT-Large backbone rather than from the proposed diffusion decoder and prototype losses. The ViT-Base ablation (54.47 mIoU) versus the ViT-Large main result (56.68 mIoU) confirms a non-negligible backbone effect, so the state-of-the-art claim requires same-backbone baselines or an explicit backbone-isolation ablation.
[Section 3.1] The MOSXAV dataset description contains an internal inconsistency. The text states that each video contains 33-70 frames and that dense annotations are provided every 5 frames across 50 sequences, totaling 2,335 frames; this implies about 46.7 annotated frames per sequence, which cannot be reconciled with 'every 5 frames' for the stated sequence lengths. Please specify the exact annotation density, the selection of the 50 sequences, and how the 488 test frames are drawn from the remaining sequences, and state the number of classes and the annotation/QA protocol for the benchmark.
[Section 3.1, Table 1] The MOSXAV evaluation does not actually demonstrate the semi-supervised setting. Table 1 reports only a GT column for MOSXAV val and test, with no Pseudo or GT+Pseudo columns, and Section 3.2 gives no pseudo-label protocol for MOSXAV. The paper's headline contribution is robustness to noisy pseudo-labels in semi-supervised learning, but on the new benchmark the model is only compared under full supervision. Please add the MOSXAV semi-supervised experiments, or explicitly restrict the semi-supervised claim to Endoscapes2023.
[Section 2.3] The prototype update mechanism is not specified. Algorithm 1 calls online_cluster(z_latent, gts, p), but the paper never states how p is initialized, updated, or normalized, nor how empty prototypes are handled. Because the method's novelty is the use of non-learnable prototype anchors, omitting this update rule prevents reproduction and leaves the stability of the online clustering unsupported.

minor comments (5)

[Section 2.3] The subsection headings appear swapped: Eq. (2) is a contrastive loss that pushes negative prototypes apart (inter-class separation), while Eq. (3) pulls pixels toward their assigned prototype (intra-class compactness). Please rename the headings accordingly.
[Section 2.1] The reverse-process description says the denoising network is applied to z0 during inference; it should be applied to the current noisy state z_t, with z0 being the predicted target.
[Algorithm 2] In Algorithm 2, the time update t_next = max(1 - (1 + step + t_diff)/steps, 0) appears to skip two diffusion intervals when t_diff=1; please clarify the units of t_diff and check the schedule against DDP's asymmetric sampling.
[Tables 1 and 2] No error bars or multiple-seed results are reported; given the limited labeled data and the noisy-pseudo-label setting, at least three runs with mean ± standard deviation should be provided for the headline numbers.
[Figure 1] Figure 1 is hard to read at the current resolution; please enlarge it and label the prototype update and contrastive-loss blocks more clearly.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the method's losses and evaluation are self-contained; the main weakness is an uncontrolled comparison, which is an experimental validity concern rather than circularity.

full rationale

I walked the paper's derivation chain. The proposed framework combines a visual encoder, a label encoder, a diffusion decoder, and prototype-anchored contrastive losses. The prototypes are obtained by online clustering over current-batch latent embeddings, and the training objective directly optimizes inter-class and intra-class contrastive terms (Eqs. 2 and 3). No output quantity is defined in terms of the predicted quantity or fitted from the test quantity. The pseudo-labels used for semi-supervised training are produced by external tools (MedSAM and SAMRefiner) rather than by the proposed model, and the evaluation is on Endoscapes2023 plus a newly introduced MOSXAV dataset whose ground truth is manually annotated by radiologists; introducing a benchmark and then evaluating on it is not a circular reduction. The only self-citations (refs. 17-19) appear in the introduction as routine examples of learning-based approaches and datasets, and none is load-bearing for the method or for the central comparison claim. The paper's principal weakness is experimental: Table 1 does not report baseline backbone capacities, input resolutions, or pseudo-label protocols, so the +5.15 mIoU margin over MedSegDiff may be confounded by the ViT-Large backbone. That is a comparability and correctness concern, not a circularity concern, and under the specified review rules it does not raise the circularity score. I therefore find no significant circularity.

Assumptions & free parameters 6 free parameters · 4 assumptions · 0 invented entities

The empirical claims rest on standard diffusion assumptions, the representativeness of the MedSAM/SAMRefiner pseudo-labels, comparability of baselines, and stability of online prototype updates. No new physical or ontological entities are introduced; MOSXAV is a dataset, not an invented entity in the sense of this ledger.

free parameters (6)

inter_loss weight = 1e-2
Set by hand in Algorithm 1; no sensitivity analysis is reported, and it controls the contribution of the contrastive loss.
intra_loss weight = 1e-3
Set by hand in Algorithm 1; no sensitivity analysis is reported.
prototype number K = 10
Selected from the ablation in Table 2b; the effect on mIoU is small (54.29 to 54.47), suggesting weak sensitivity.
temperature tau = not reported
Equation 2 depends on tau for contrastive loss concentration; the value is not given in the paper.
label scaling factor s = not reported
Controls the signal-to-noise ratio and denoising difficulty in Section 2.2; the value is not specified.
inference time interval t_diff = 1
Set empirically in Algorithm 2; no ablation is provided for this choice.

assumptions (4)

standard math Cosine noise schedule and DDIM update rule are valid for discrete label embeddings.
Section 2.1 and 2.4 invoke the standard diffusion forward and reverse process without proof; this is accepted background in the field.
domain assumption Pseudo-labels produced by MedSAM and SAMRefiner on the BBox201 subset are a realistic surrogate for noisy pseudo-labels in semi-supervised learning.
Section 3.1 describes the pseudo-label pipeline; the claimed robustness to noisy pseudo-labels depends on these labels being noisy in a representative way.
domain assumption Baseline methods in Table 1 were trained under a comparable protocol, backbone capacity, and data.
The state-of-the-art claim requires fair comparison; the paper does not report baseline training details.
ad hoc to paper Online clustering prototype updates remain stable without gradient-based learning.
Algorithm 1 updates prototypes via online_cluster each batch, yet the text calls them non-learnable fixed anchors; stability is assumed without analysis.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Robust Noisy Pseudo-label Learning for Semi-supervised Medical Image Segmentation Using Diffusion Model." pith.science (2026). https://pith.science/paper/NCV5EXCU

@misc{pith2026250716429,
  author       = {Pith},
  title        = {Pith review of: Robust Noisy Pseudo-label Learning for Semi-supervised Medical Image Segmentation Using Diffusion Model},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/NCV5EXCU}},
  note         = {Machine review of arXiv:2507.16429}
}

read the original abstract

Obtaining pixel-level annotations in the medical domain is both expensive and time-consuming, often requiring close collaboration between clinical experts and developers. Semi-supervised medical image segmentation aims to leverage limited annotated data alongside abundant unlabeled data to achieve accurate segmentation. However, existing semi-supervised methods often struggle to structure semantic distributions in the latent space due to noise introduced by pseudo-labels. In this paper, we propose a novel diffusion-based framework for semi-supervised medical image segmentation. Our method introduces a constraint into the latent structure of semantic labels during the denoising diffusion process by enforcing prototype-based contrastive consistency. Rather than explicitly delineating semantic boundaries, the model leverages class prototypes centralized semantic representations in the latent space as anchors. This strategy improves the robustness of dense predictions, particularly in the presence of noisy pseudo-labels. We also introduce a new publicly available benchmark: Multi-Object Segmentation in X-ray Angiography Videos (MOSXAV), which provides detailed, manually annotated segmentation ground truth for multiple anatomical structures in X-ray angiography videos. Extensive experiments on the EndoScapes2023 and MOSXAV datasets demonstrate that our method outperforms state-of-the-art medical image segmentation approaches under the semi-supervised learning setting. This work presents a robust and data-efficient diffusion model that offers enhanced flexibility and strong potential for a wide range of clinical applications.

Figures

Figures reproduced from arXiv: 2507.16429 by the authors.

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

35 extracted references · 31 canonical work pages

[1]

In: Proceedings of the European Conference on Computer Vision

Cao, H., et al.: Swin-unet: Unet-like pure transformer for medical image segmenta- tion. In: Proceedings of the European Conference on Computer Vision. pp. 205–218 (2022) 2, 8

work page 2022
[2]

arXiv preprint arXiv:2102.04306 (2021) 8

Chen, J., et al.: Transunet: Transformers make strong encoders for medical image segmentation. arXiv preprint arXiv:2102.04306 (2021) 8

arXiv 2021
[3]

In: Proceedings of the International Conference on Machine Learning

Chen, T., et al.: A simple framework for contrastive learning of visual represen- tations. In: Proceedings of the International Conference on Machine Learning. vol. 119, pp. 1597–1607 (2020) 5

work page 2020
[4]

In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition

Chen, X., et al.: Semi-supervised semantic segmentation with cross pseudo super- vision. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. pp. 2613–2622 (2021) 2

work page 2021
[5]

In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition

Chi, H., et al.: Adaptive bidirectional displacement for semi-supervised medical image segmentation. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. pp. 4070–4080 (2024) 2

work page 2024
[6]

In: International Conference on Learning Representations (2021) 4, 8

Dosovitskiy, A., et al.: An image is worth 16x16 words: Transformers for image recognition at scale. In: International Conference on Learning Representations (2021) 4, 8

work page 2021
[7]

In: Advances in Neural Information Processing Systems

Ho, J., et al.: Denoising diffusion probabilistic models. In: Advances in Neural Information Processing Systems. vol. 33, pp. 6840–6851 (2020) 3

work page 2020
[8]

arXiv preprint arXiv:2012.12453 (2020) 2

Hong, W.Y., et al.: Cholecseg8k: a semantic segmentation dataset for laparoscopic cholecystectomy based on cholec80. arXiv preprint arXiv:2012.12453 (2020) 2

arXiv 2020

Show all 35 references

[9]

In: Proceedings of the IEEE International Conference on Computer Vision

Ji, Y., et al.: Ddp: Diffusion model for dense visual prediction. In: Proceedings of the IEEE International Conference on Computer Vision. pp. 21684–21695 (2023) 2, 5, 6

2023
[10]

Expert Systems41(12), e13708 (2024) 7

Jiménez-Partinen, A., et al.: Cadica: A new dataset for coronary artery disease detection by using invasive coronary angiography. Expert Systems41(12), e13708 (2024) 7

2024
[11]

In: Proceedings of the AAAI Conference on Artificial Intelligence

Le, M.Q., et al.: Maskdiff: Modeling mask distribution with diffusion probabilistic model for few-shot instance segmentation. In: Proceedings of the AAAI Conference on Artificial Intelligence. vol. 38, pp. 2874–2881 (2024) 2

2024
[12]

In: International Conference on Learning Representations (2025) 8

Lin, Y., et al.: Samrefiner: Taming segment anything model for universal mask refinement. In: International Conference on Learning Representations (2025) 8

2025
[13]

In: Proceedings of Medical Image Computing and Computer-Assisted Intervention (2024) 2

Liu, X., et al.: Diffrect: Latent diffusion label rectification for semi-supervised medical image segmentation. In: Proceedings of Medical Image Computing and Computer-Assisted Intervention (2024) 2

2024
[14]

Medical Physics49(2), 1262–1275 (2022) 2

Luo, Y., et al.: Edge-enhancement densenet for x-ray fluoroscopy image denoising in cardiac electrophysiology procedures. Medical Physics49(2), 1262–1275 (2022) 2

2022
[15]

Ma, J., et al.: Abdomenct-1k: Is abdominal organ segmentation a solved problem? IEEE Transactions on Pattern Analysis and Machine Intelligence44(10), 6695– 6714 (2022) 2

2022
[16]

Nature Communications15(1), 654 (2024) 2, 8 Robust Noisy Pseudo-label Learning Using Diffusion Model 11

Ma, J., et al.: Segment anything in medical images. Nature Communications15(1), 654 (2024) 2, 8 Robust Noisy Pseudo-label Learning Using Diffusion Model 11

2024
[17]

Medical physics45(11), 5066–5079 (2018) 2

Ma, Y., et al.: A novel real-time computational framework for detecting catheters and rigid guidewires in cardiac catheterization procedures. Medical physics45(11), 5066–5079 (2018) 2

2018
[18]

Physics in Medicine & Biology66(5), 055019 (2021) 2

Ma, Y., et al.: Real-time registration of 3d echo to x-ray fluoroscopy based on cascading classifiers and image registration. Physics in Medicine & Biology66(5), 055019 (2021) 2

2021
[19]

IEEE Transactions on Biomedical Engineering69(2), 635–644 (2022) 2

Ma, Y., et al.: A tensor-based catheter and wire detection and tracking framework and its clinical applications. IEEE Transactions on Biomedical Engineering69(2), 635–644 (2022) 2

2022
[20]

IEEE Transactions on Medical Imaging34(10), 1993–2024 (2015) 2

Menze, B.H., et al.: The multimodal brain tumor image segmentation benchmark. IEEE Transactions on Medical Imaging34(10), 1993–2024 (2015) 2

2015
[21]

In: Proceedings of Medical Image Computing and Computer-Assisted Intervention

Miao, J., et al.: Cross prompting consistency with segment anything model for semi-supervised medical image segmentation. In: Proceedings of Medical Image Computing and Computer-Assisted Intervention. vol. LNCS 15011 (2024) 2

2024
[22]

arXiv preprint arXiv:2312.12429 (2023) 7

Murali, A., et al.: The endoscapes dataset for surgical scene segmentation, object detection, and critical view of safety assessment: Official splits and benchmark. arXiv preprint arXiv:2312.12429 (2023) 7

2023 arXiv
[23]

In: Proceed- ings of the International Conference on Machine Learning

Nichol, A.Q., et al.: Improved denoising diffusion probabilistic models. In: Proceed- ings of the International Conference on Machine Learning. vol. 139, pp. 8162–8171 (2021) 5

2021
[24]

arXiv preprint arXiv:1804.03999 (2018) 2, 7, 8

Oktay, O., et al.: Attention u-net: Learning where to look for the pancreas. arXiv preprint arXiv:1804.03999 (2018) 2, 7, 8

2018 arXiv
[25]

In: Medical Image Computing and Computer-Assisted Intervention

Ronneberger, O., et al.: U-net: Convolutional networks for biomedical image seg- mentation. In: Medical Image Computing and Computer-Assisted Intervention. pp. 234–241 (2015) 2, 7, 8

2015
[26]

In: Proceedings of the International Conference on Machine Learning

Sohl-Dickstein, J., et al.: Deep unsupervised learning using nonequilibrium thermo- dynamics. In: Proceedings of the International Conference on Machine Learning. vol. 37, pp. 2256–2265 (2015) 3

2015
[27]

In: Advances in Neural Information Processing Systems

Sohn, K., et al.: Fixmatch: Simplifying semi-supervised learning with consistency and confidence. In: Advances in Neural Information Processing Systems. vol. 33, pp. 596–608 (2020) 2

2020
[28]

In: International Conference on Learning Representations (2021) 3, 6

Song, J., et al.: Denoising diffusion implicit models. In: International Conference on Learning Representations (2021) 3, 6

2021
[29]

In: Proceedings of the IEEE International Symposium on Biomedical Imaging

Tang, F., et al.: Cmu-net: A strong convmixer-based medical ultrasound image segmentation network. In: Proceedings of the IEEE International Symposium on Biomedical Imaging. pp. 1–5 (2023) 8

2023
[30]

In: Proceedings of the IEEE International Sympo- sium on Biomedical Imaging

Tang, F., et al.: Cmunext: An efficient medical image segmentation network based on large kernel and skip fusion. In: Proceedings of the IEEE International Sympo- sium on Biomedical Imaging. pp. 1–5 (2024) 8

2024
[31]

arXiv preprint arXiv:2408.03393 (2024) 2

Wahid, F., et al.: Biomedical image segmentation: a systematic literature review of deep learning based object detection methods. arXiv preprint arXiv:2408.03393 (2024) 2

2024 arXiv
[32]

In: Proceedings of the AAAI Conference on Artificial Intelligence

Wu, J., et al.: Medsegdiff-v2: Diffusion-based medical image segmentation with transformer. In: Proceedings of the AAAI Conference on Artificial Intelligence. vol. 38, pp. 6030–6038 (2024) 2, 8

2024
[33]

Pattern Recognition166, 111722 (2025) 2

Zhang, H., et al.: Prototype-augmented mean teacher for robust semi-supervised medical image segmentation. Pattern Recognition166, 111722 (2025) 2

2025
[34]

In: Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining

Zhang, X., et al.: Adversarial variational embedding for robust semi-supervised learning. In: Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining. p. 139–147 (2019) 2 12 L. Xi et al

2019
[35]

In: Pro- ceedings of Medical Image Computing and Computer-Assisted Intervention

Zhang, Y., et al.: Generator versus segmentor: Pseudo-healthy synthesis. In: Pro- ceedings of Medical Image Computing and Computer-Assisted Intervention. pp. 150–160 (2021) 2

2021

Pith tools

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

[1] [1]

In: Proceedings of the European Conference on Computer Vision

Cao, H., et al.: Swin-unet: Unet-like pure transformer for medical image segmenta- tion. In: Proceedings of the European Conference on Computer Vision. pp. 205–218 (2022) 2, 8

work page 2022

[2] [2]

arXiv preprint arXiv:2102.04306 (2021) 8

Chen, J., et al.: Transunet: Transformers make strong encoders for medical image segmentation. arXiv preprint arXiv:2102.04306 (2021) 8

arXiv 2021

[3] [3]

In: Proceedings of the International Conference on Machine Learning

Chen, T., et al.: A simple framework for contrastive learning of visual represen- tations. In: Proceedings of the International Conference on Machine Learning. vol. 119, pp. 1597–1607 (2020) 5

work page 2020

[4] [4]

In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition

Chen, X., et al.: Semi-supervised semantic segmentation with cross pseudo super- vision. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. pp. 2613–2622 (2021) 2

work page 2021

[5] [5]

In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition

Chi, H., et al.: Adaptive bidirectional displacement for semi-supervised medical image segmentation. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. pp. 4070–4080 (2024) 2

work page 2024

[6] [6]

In: International Conference on Learning Representations (2021) 4, 8

Dosovitskiy, A., et al.: An image is worth 16x16 words: Transformers for image recognition at scale. In: International Conference on Learning Representations (2021) 4, 8

work page 2021

[7] [7]

In: Advances in Neural Information Processing Systems

Ho, J., et al.: Denoising diffusion probabilistic models. In: Advances in Neural Information Processing Systems. vol. 33, pp. 6840–6851 (2020) 3

work page 2020

[8] [8]

arXiv preprint arXiv:2012.12453 (2020) 2

Hong, W.Y., et al.: Cholecseg8k: a semantic segmentation dataset for laparoscopic cholecystectomy based on cholec80. arXiv preprint arXiv:2012.12453 (2020) 2

arXiv 2020