Pith. sign in

REVIEW 5 major objections 5 minor 3 cited by

GraphCL: Graph-based Clustering for Semi-Supervised Medical Image Segmentation

T0 review · 5 major / 5 minor · reviewed 2026-08-12 · deepseek-v4-flash

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

Pith's one-line read The paper claims that modeling medical images as graphs and adding a correlation-clustering loss improves semi-supervised segmentation, reporting top Dice scores on LA, ACDC, and Pancreas-NIH.

desk verdict A plausible incremental method for semi-supervised medical segmentation whose central empirical claim is undermined by an internal contradiction in the ACDC ablation. read the letter →

arxiv 2411.13147 v2 pith:FCKPGN6K submitted 2024-11-20 cs.CV cs.AI

Mengzhu Wang , Jiao Li , Houcheng Su , Nan Yin , Liang Yang , Shen Li This is my paper · ORCID
classification cs.CVcs.AI
keywords semi-supervisedlearningmedicalimagesegmentationgraphneuralnetworkscorrelationclusteringconvolutionalteacher-studentframeworkvolumetricstructure-awarealignment
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 tries to establish that semi-supervised medical image segmentation improves when the model explicitly represents graph structure and clusters features with a graph neural network. It argues that existing methods lean on complex training strategies while ignoring structural relationships between samples and between voxels. GraphCL builds a dense instance graph from CNN features, propagates information through a graph convolutional network, and adds a correlation-clustering loss that groups similar voxels without needing a preset number of clusters. On three benchmarks (LA, ACDC, Pancreas-NIH) with 5–20% labeled volumes, it reports higher overlap-based Dice and Jaccard scores and generally lower boundary distances than the compared semi-supervised methods. If the claim holds, adding graph-structure modeling to teacher-student pipelines is a simple and effective way to use unlabeled data.

What carries the argument

The load-bearing object is a graph built from the network's own features. Each sample in a mini-batch becomes a node, and a Data Structure Analyzer produces structure scores $X_{sa}$; the adjacency matrix is $\hat{A} = X_{sa}X_{sa}^\top$, on which a graph convolutional network propagates features as $Z = \hat{D}^{-1/2}\hat{A}\hat{D}^{-1/2}XW$. A second, voxel-level graph uses the correlation of deep features, $W = F F^\top - \max(F F^\top)/\tau$, and a GNN plus an MLP with softmax produces the cluster assignment matrix $S$. The correlation-clustering loss $L_{CC} = -\operatorname{Tr}(W S S^\top)$, added to the segmentation loss with weight $\kappa$, is the mechanism intended to make similar voxels cluster together without a predefined number of clusters.

What would settle it

Fix everything in the pipeline but replace the Data Structure Analyzer's structure scores with random noise; if the reported gains persist, the structure-aware graph is not doing the claimed work. In the same spirit, inspect the learned cluster assignments $S$ on a validation fold: if voxels from different organs fall mostly into one or two clusters, the clustering objective has collapsed and the reported gains must come from the copy-paste baseline.

Watch

Extended reading notes

Core claim

GraphCL's central claim is that jointly modeling data structure—at the image level through a structure-aware graph and at the voxel level through graph clustering—improves semi-supervised segmentation in one unified network. The paper says this is the first work to model data-structure information for semi-supervised medical image segmentation. On the LA and ACDC datasets, it claims to outperform the compared state-of-the-art methods on all four reported metrics at both 5% and 10% labeled ratios; on Pancreas-NIH, it reports higher Dice, Jaccard, and 95HD than the best compared method. The reported best Dice scores are 88.80% and 90.24% on LA, 88.68% and 89.31% on ACDC, and 83.15% on Pancreas-NIH. The driving objective is the correlation-clustering loss $L_{CC} = -\operatorname{Tr}(W S S^\top)$, which rewards giving similar voxels the same cluster assignment while separating dissimilar voxels.

Load-bearing premise

Everything rests on the graph built from the network's own features capturing true anatomical structure, and on the clustering loss pushing features into useful groups instead of collapsing to trivial assignments.

Editorial extensions

If this is right

  • If GraphCL's empirical claim is right, graph-structure modeling is a viable source of improvement for semi-supervised medical segmentation, not just for supervised or weakly supervised tasks.
  • The k-less correlation-clustering loss can act as a regularizer that sharpens boundary delineation, according to the paper's ablation studies, without requiring the user to specify the number of clusters.
  • The reported gains are largest at the 5% labeled ratio on LA and ACDC, suggesting graph structure is most helpful when labeled data are scarcest.
  • The ablation over GCN placement implies the position of the graph module inside the encoder matters, with deeper insertion giving the best results.

Reading between the lines

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

  • The paper leaves implicit that the clustering loss could be tested as a standalone regularizer on other semi-supervised baselines; if it helps them too, the benefit is generic rather than tied to the specific graph construction.
  • Because the adjacency matrix and cluster assignments are computed from the same features the loss optimizes, a direct check of cluster purity against ground-truth anatomy would show whether $S$ actually corresponds to organs or is merely a feature regularizer.
  • The ASD result on Pancreas-NIH, where GraphCL does not beat the best baseline on average surface distance, suggests that boundary-distance behavior is not uniformly improved; a natural follow-up is to weight the clustering loss per class or combine it with an explicit boundary term.
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 / 5 minor

Summary. The paper proposes GraphCL, a semi-supervised medical image segmentation method built on the BCP teacher-student framework. It adds two graph-based components: a structure-aware alignment module that constructs an instance graph from DSA-derived structure scores and propagates features with a GCN, and a correlation clustering loss L_CC computed from a patch-affinity matrix. The method is evaluated on LA, ACDC, and Pancreas-NIH with Dice, Jaccard, 95HD, and ASD, reporting improvements over BCP and other baselines in most settings.

Significance. The direction is plausible: using graph structure and clustering as a regularizer for SSMIS is a reasonable extension of the BCP line, and the reported Dice/Jaccard gains are consistently positive across three datasets. However, the paper's central claim is empirical, and the current evidence is weakened by internal inconsistencies in the ablations, an implausible table entry, incorrect text-vs-table reporting, and the absence of variance information and code. The contribution is potentially useful if those issues are resolved.

major comments (5)
  1. [4.4, Table 4] In Table 4, on ACDC with 10% labeled data, the full model (SA+LCC) reaches Dice 89.31, while both SA-only (89.52) and LCC-only (89.53) are higher. This directly contradicts the Section 4.4 statement that 'incorporating both SA and LCC achieves the best results,' and it undermines the component-combination claim that supports the method. Additionally, on LA with 10% labeled data, LCC-only (88.79) is below the baseline (89.39), which contradicts the text's claim that adding SA or LCC individually improves performance over the baseline. The authors should report results with error bars or multiple seeds and explain these discrepancies.
  2. [4.3, ACDC paragraph] Section 4.3 states that on ACDC with 10% labeled data GraphCL outperforms the second-best approach by an average of 26.01% across all four metrics, and by 29.65% with 5% labeled data. These numbers do not match Table 2, where absolute Dice/Jaccard differences over BCP are below 3 percentage points and 95HD/ASD are measured in distance units. The average-improvement figures need to be recalculated or removed.
  3. [4.4, Table 6] Table 6 reports Dice of 0.030 and 0.008 with Jaccard of 0.016 and 0.004 for the '1 layer' configuration on the ACDC dataset, along with 95HD of 80.51 and 40.94. These values are implausible for the evaluated segmentation task and are not discussed in the text. Because this table is used to support the claim that deeper GCN placement is optimal, it cannot serve as evidence in its current form.
  4. [3.3.1-3.3.2] The Data Structure Analyzer is only characterized by its output X_sa used in Eq. (13); its architecture, input, and output dimension h are never specified. Similarly, the 'k-less' correlation clustering in Eq. (17) uses a softmax MLP whose output dimension implicitly defines the number of clusters, but that dimension and the handling of the softmax assignments are not given. Without these details, the method is not reproducible, and the risk that L_CC in Eq. (18) degenerates to trivial cluster assignments cannot be assessed.
  5. [Abstract; Table 3] The abstract states that GraphCL outperforms state-of-the-art semi-supervised medical image segmentation methods, but on Pancreas-NIH Table 3 shows ASD of 2.12, which is worse than CoraNet's 1.89. The claim should be qualified per metric and per dataset. In addition, no error bars or multiple-seed statistics are provided anywhere, and no code is released, so the robustness of the headline improvements cannot be assessed.
minor comments (5)
  1. [Eq. (15)] The expression W = F·F^T - Max(F·F^T) / τ is ambiguous; parentheses should clarify whether the maximum is subtracted before or after division by τ, and the role of τ as a temperature should be stated consistently with Section 4.5.
  2. [Section 4.5] The text says κ controls the weight of the structure-aware alignment, but Eq. (19) multiplies κ by the clustering loss L_CC. This inconsistency should be corrected.
  3. [Figure 5] The visualization is described as 'kernel dense estimations' but the axes, color scale, and the feature layer used are not identified, making it hard to interpret.
  4. [Introduction] The claim of being the first to model data structure information for SSMIS is not substantiated; given the graph-based SSL surveys and GNN segmentation work cited in Section 2, the novelty claim should be softened or placed in a more precise context.
  5. [References] Some references contain typos (e.g., 'NeruIPS' should be 'NeurIPS'); please proofread the reference list.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: GraphCL is an empirical method paper whose central claim is benchmarked against external baselines, and its clustering objective is a self-supervised regularizer rather than a fitted prediction.

full rationale

The paper makes no derivation-level claim that would reduce to its own inputs. The adjacency matrix A_hat = X_sa X_sa^T is learned from CNN features and used for GCN propagation; the clustering loss L_CC = -Tr(W S S^T) is a training regularizer that encourages cluster assignments S to match pairwise affinities W derived from the same features. This is a self-supervised objective, not a circular validation of the central claim, because the reported superiority over SOTA is measured on held-out test sets of standard benchmarks (LA, ACDC, Pancreas-NIH) against external baselines. No fitted parameter is renamed as a prediction, and no load-bearing argument depends on self-citation. The ablation inconsistency on ACDC 10% (full model 89.31 Dice below SA-only 89.52 and LCC-only 89.53) and the implausible Table 6 entries are correctness and reproducibility concerns, but they are not instances of circularity under the definitions used here.

Assumptions & free parameters 6 free parameters · 3 assumptions · 1 invented entities

The central claim is empirical, so the ledger mainly tracks hyperparameters and modeling assumptions. Key free parameters are the loss weights (alpha, kappa), the clustering sensitivity tau, and the unreported DSA output dimension and number of clusters. The key axioms are the reliability of teacher pseudo-labels, the informativeness of the constructed graphs, and the non-degeneracy of the correlation clustering objective.

free parameters (6)
  • alpha = 0.5
    Weight for unlabeled pixel loss in the bidirectional copy-paste loss (Eq. 7-8); chosen without sensitivity analysis.
  • kappa = 0.01
    Weight for the graph clustering loss L_CC (Eq. 19); selected via sensitivity analysis on ACDC (Section 4.5).
  • tau = 2
    Clustering sensitivity parameter in the affinity matrix W (Eq. 15); selected via sensitivity analysis on ACDC (Section 4.5).
  • DSA output dimension h = not reported
    Dimension of structure scores X_sa used to compute the graph adjacency matrix (Eq. 13); not specified in the paper.
  • number of clusters K = not reported
    Output dimension of the MLP in Eq. 17, which determines the number of clusters for the assignment matrix S; not specified anywhere.
  • GCN layer placement = fifth layer
    Ablation in Section 4.4 shows the fifth decoder layer works best; the default is not stated in the implementation details.
assumptions (3)
  • domain assumption The teacher-student EMA framework (Eq. 11) provides reliable pseudo-labels for unlabeled images.
    Standard SSL assumption invoked in Section 3.2; no analysis of pseudo-label noise is provided.
  • ad hoc to paper Graph convolution (Eq. 14) using adjacency A_hat = X_sa X_sa^T captures structural relationships useful for segmentation.
    No theoretical or empirical justification beyond final metrics; the DSA producing X_sa is underspecified.
  • ad hoc to paper The correlation clustering objective L_CC = -Tr(W S S^T) encourages semantically meaningful clusters when optimized with soft assignments.
    The paper asserts this in Section 3.3.2 but provides no analysis; the ablation shows LCC alone can hurt LA performance (Table 4).
invented entities (1)
  • Data Structure Analyzer (DSA)
    purpose: Generates structure scores X_sa used to construct the image-level graph adjacency matrix (Eq. 13).
    Introduced as a new module but with no architectural specification, training details, or independent validation; it is only mentioned in Section 3.3.1.

how reviews work

0 comments
Cite this review

Pith. "Pith review of GraphCL: Graph-based Clustering for Semi-Supervised Medical Image Segmentation." pith.science (2026). https://pith.science/paper/FCKPGN6K

@misc{pith2026241113147,
  author       = {Pith},
  title        = {Pith review of: GraphCL: Graph-based Clustering for Semi-Supervised Medical Image Segmentation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/FCKPGN6K}},
  note         = {Machine review of arXiv:2411.13147}
}
read the original abstract

Semi-supervised learning (SSL) has made notable advancements in medical image segmentation (MIS), particularly in scenarios with limited labeled data and significantly enhancing data utilization efficiency. Previous methods primarily focus on complex training strategies to utilize unlabeled data but neglect the importance of graph structural information. Different from existing methods, we propose a graph-based clustering for semi-supervised medical image segmentation (GraphCL) by jointly modeling graph data structure in a unified deep model. The proposed GraphCL model enjoys several advantages. Firstly, to the best of our knowledge, this is the first work to model the data structure information for semi-supervised medical image segmentation (SSMIS). Secondly, to get the clustered features across different graphs, we integrate both pairwise affinities between local image features and raw features as inputs. Extensive experimental results on three standard benchmarks show that the proposed GraphCL algorithm outperforms state-of-the-art semi-supervised medical image segmentation methods.

Figures

Figures reproduced from arXiv: 2411.13147 by the authors.

Figure 1
Figure 1. We apply graph neural networks (GNN) to address [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. The proposed Graph-based Clustering for Semi-supervised Medical Image Segmentation (GraphCL) architecture consists of [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Sensitivity analysis on the ACDC dataset with the [PITH_FULL_IMAGE:figures/full_fig_p009_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Sensitivity analysis on the ACDC dataset with the [PITH_FULL_IMAGE:figures/full_fig_p009_4.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 3 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. UniCT Depth: Event-Image Fusion Based Monocular Depth Estimation with Convolution-Compensated ViT Dual SA Block

    cs.CV 2025-07 conditional novelty 4.0 of 10

    UniCT Depth fuses event and image data with a CNN-Transformer hybrid to improve monocular depth estimation, achieving lower average errors on MVSEC and DENSE benchmarks.

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

    eess.IV 2025-05 conditional novelty 4.0 of 10

    TransMedSeg applies ISDA-style feature augmentation with teacher-student class statistics to semi-supervised medical segmentation, reporting marginal SOTA improvements on ACDC, LA, and Pancreas-NIH.

  3. In-context learning for medical image segmentation

    eess.IV 2024-12 conditional novelty 4.0 of 10

    A cascade that reuses each predicted slice as a new support example improves few-shot sequential medical image segmentation over plain UniverSeg on 5 of 8 cardiac structures.

Reference graph

Works this paper leans on

57 extracted references · 56 canonical work pages · cited by 3 Pith papers

  1. [1]

    Slic superpixels compared to state-of-the-art superpixel methods

    Radhakrishna Achanta, Appu Shaji, Kevin Smith, Aurelien Lucchi, Pascal Fua, and Sabine S ¨usstrunk. Slic superpixels compared to state-of-the-art superpixel methods. TPAMI, 34 (11):2274–2282, 2012. 2

    work page 2012

  2. [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. In CVPR, pages 11514–11524, 2023. 1, 3, 6, 7

    work page 2023

  3. [3]

    Semi-supervised clustering methods

    Eric Bair. Semi-supervised clustering methods. Wiley In- terdisciplinary Reviews: Computational Statistics, 5(5):349– 361, 2013. 2

    work page 2013

  4. [4]

    Deep learning techniques for automatic mri cardiac multi-structures segmentation and diagnosis: is the problem solved? TMI, 37(11):2514–2525, 2018

    Olivier Bernard, Alain Lalande, Clement Zotti, Freder- ick Cervenansky, Xin Yang, Pheng-Ann Heng, Irem Cetin, Karim Lekadir, Oscar Camara, Miguel Angel Gonzalez Ballester, et al. Deep learning techniques for automatic mri cardiac multi-structures segmentation and diagnosis: is the problem solved? TMI, 37(11):2514–2525, 2018. 6

    work page 2018

  5. [5]

    Computed tomography

    Thorsten M Buzug. Computed tomography. In Springer Handbook of Medical Technology, pages 311–342. 2011. 1

    work page 2011

  6. [6]

    Adaptive bidirectional displacement for semi-supervised medical image segmentation

    Hanyang Chi, Jian Pang, Bingfeng Zhang, and Weifeng Liu. Adaptive bidirectional displacement for semi-supervised medical image segmentation. In CVPR, pages 4070–4080,

    work page

  7. [7]

    Graph-based semi-supervised learning: A review

    Yanwen Chong, Yun Ding, Qing Yan, and Shaoming Pan. Graph-based semi-supervised learning: A review. Neuro- computing, 408:216–230, 2020. 2

    work page 2020

  8. [8]

    Domain adaptation for med- ical image segmentation using transformation-invariant self- training

    Negin Ghamsarian, Javier Gamazo Tejero, Pablo M ´arquez- Neila, Sebastian Wolf, Martin Zinkernagel, Klaus Schoeff- mann, and Raphael Sznitman. Domain adaptation for med- ical image segmentation using transformation-invariant self- training. In MICCAI, pages 331–341, 2023. 2

    work page 2023

Show all 57 references
  1. [9]

    Overview of functional magnetic resonance imaging

    Gary H Glover. Overview of functional magnetic resonance imaging. Neurosurgery Clinics, 22(2):133–139, 2011. 1

    2011

  2. [10]

    A new model for learning in graph domains

    Marco Gori, Gabriele Monfardini, and Franco Scarselli. A new model for learning in graph domains. In IJCNN, pages 729–734, 2005. 3

    2005

  3. [11]

    Unsupervised clustering using pseudo- semi-supervised learning

    Divam Gupta, Ramachandran Ramjee, Nipun Kwatra, and Muthian Sivathanu. Unsupervised clustering using pseudo- semi-supervised learning. In ICLR, 2020. 2

    2020

  4. [12]

    Semicvt: Semi-supervised convolutional vi- sion transformer for semantic segmentation

    Huimin Huang, Shiao Xie, Lanfen Lin, Ruofeng Tong, Yen- Wei Chen, Yuexiang Li, Hong Wang, Yawen Huang, and Yefeng Zheng. Semicvt: Semi-supervised convolutional vi- sion transformer for semantic segmentation. In CVPR, pages 11340–11349, 2023. 4

    2023

  5. [13]

    Detection recovery in online multi-object track- ing with sparse graph tracker

    Jeongseok Hyun, Myunggu Kang, Dongyoon Wee, and Dit- Yan Yeung. Detection recovery in online multi-object track- ing with sparse graph tracker. In WACV, pages 4850–4859,

  6. [14]

    Adam: A method for stochastic opti- mization

    Diederik P Kingma. Adam: A method for stochastic opti- mization. arXiv preprint arXiv:1412.6980, 2014. 6

    2014 arXiv

  7. [15]

    Semi-supervised classifi- cation with graph convolutional networks

    Thomas N Kipf and Max Welling. Semi-supervised classifi- cation with graph convolutional networks. ICLR, 2016. 2

    2016

  8. [16]

    Distance metric learning using graph convolutional networks: Appli- cation to functional brain networks

    Sofia Ira Ktena, Sarah Parisot, Enzo Ferrante, Martin Rajchl, Matthew Lee, Ben Glocker, and Daniel Rueckert. Distance metric learning using graph convolutional networks: Appli- cation to functional brain networks. In MICCAI, pages 469– 477, 2017. 3

    2017

  9. [17]

    Graph networks for multiple object tracking

    Jiahe Li, Xu Gao, and Tingting Jiang. Graph networks for multiple object tracking. In WACV, pages 719–728, 2020. 2

    2020

  10. [18]

    Shape-aware semi-supervised 3d semantic segmentation for medical im- ages

    Shuailin Li, Chuyu Zhang, and Xuming He. Shape-aware semi-supervised 3d semantic segmentation for medical im- ages. In MICCAI, pages 552–561, 2020. 6, 7

    2020

  11. [19]

    Semi-supervised clustering in attributed het- erogeneous information networks

    Xiang Li, Yao Wu, Martin Ester, Ben Kao, Xin Wang, and Yudian Zheng. Semi-supervised clustering in attributed het- erogeneous information networks. In WWW, pages 1621– 1629, 2017. 2

    2017

  12. [20]

    Semi-supervised clustering with deep metric learning and graph embedding

    Xiaocui Li, Hongzhi Yin, Ke Zhou, and Xiaofang Zhou. Semi-supervised clustering with deep metric learning and graph embedding. WWW, 23:781–798, 2020. 2

    2020

  13. [21]

    Towards deeper graph neural networks

    Meng Liu, Hongyang Gao, and Shuiwang Ji. Towards deeper graph neural networks. In KDD, pages 338–348, 2020. 3

    2020

  14. [22]

    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. In AAAI, pages 8801–8809, 2021. 2, 6, 7

    2021

  15. [23]

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

    Xiangde Luo, Wenjun Liao, Jieneng Chen, Tao Song, Yi- nan Chen, Shichuan Zhang, Nianyong Chen, Guotai Wang, and Shaoting Zhang. Efficient semi-supervised gross target volume of nasopharyngeal carcinoma segmentation via un- certainty rectified pyramid consistency. In MICCAI, pa...

    2021

  16. [24]

    Gcan: Graph convolutional adversarial network for unsupervised domain adaptation

    Xinhong Ma, Tianzhu Zhang, and Changsheng Xu. Gcan: Graph convolutional adversarial network for unsupervised domain adaptation. In CVPR, pages 8266–8276, 2019. 3

    2019

  17. [25]

    Demystifying struc- tural disparity in graph neural networks: Can one size fit all? NeurIPS, 36, 2024

    Haitao Mao, Zhikai Chen, Wei Jin, Haoyu Han, Yao Ma, Tong Zhao, Neil Shah, and Jiliang Tang. Demystifying struc- tural disparity in graph neural networks: Can one size fit all? NeurIPS, 36, 2024. 3

    2024

  18. [26]

    Caussl: Causality-inspired semi-supervised learning for medical image segmentation

    Juzheng Miao, Cheng Chen, Furui Liu, Hao Wei, and Pheng- Ann Heng. Caussl: Causality-inspired semi-supervised learning for medical image segmentation. In ICCV, pages 21426–21437, 2023. 3

    2023

  19. [27]

    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. In 3DV, pages 565–571, 2016. 6, 7

    2016

  20. [28]

    Graphnet: Learning image pseudo annotations for weakly- supervised semantic segmentation

    Mengyang Pu, Yaping Huang, Qingji Guan, and Qi Zou. Graphnet: Learning image pseudo annotations for weakly- supervised semantic segmentation. In ACM MM, pages 483– 491, 2018. 2

    2018

  21. [29]

    Re- search progress on semi-supervised clustering

    Yue Qin, Shifei Ding, Lijuan Wang, and Yanru Wang. Re- search progress on semi-supervised clustering. Cognitive Computation, 11(5):599–612, 2019. 2

    2019

  22. [30]

    Efficient 3d semantic segmentation with superpoint transformer

    Damien Robert, Hugo Raguet, and Loic Landrieu. Efficient 3d semantic segmentation with superpoint transformer. In ICCV, pages 17195–17204, 2023. 4

    2023

  23. [31]

    U-net: Convolutional networks for biomedical image segmentation

    Olaf Ronneberger, Philipp Fischer, and Thomas Brox. U-net: Convolutional networks for biomedical image segmentation. In MICCAI, pages 234–241, 2015. 6, 7

    2015

  24. [32]

    Deeporgan: Multi-level deep convolutional networks for automated pan- creas segmentation

    Holger R Roth, Le Lu, Amal Farag, Hoo-Chang Shin, Jiamin Liu, Evrim B Turkbey, and Ronald M Summers. Deeporgan: Multi-level deep convolutional networks for automated pan- creas segmentation. In MICCAI, pages 556–564, 2015. 6

    2015

  25. [33]

    The graph neural net- work model

    Franco Scarselli, Marco Gori, Ah Chung Tsoi, Markus Ha- genbuchner, and Gabriele Monfardini. The graph neural net- work model. IEEE Transactions on Neural Networks, 20(1): 61–80, 2008. 3

    2008

  26. [34]

    Modeling re- lational data with graph convolutional networks

    Michael Schlichtkrull, Thomas N Kipf, Peter Bloem, Rianne Van Den Berg, Ivan Titov, and Max Welling. Modeling re- lational data with graph convolutional networks. In ESWC, pages 593–607, 2018. 3

    2018

  27. [35]

    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. TMI, 41(3):608–620, 2021. 1, 2, 6, 7

    2021

  28. [36]

    Graphalign: Enhancing accurate feature alignment by graph matching for multi-modal 3d object detection

    Ziying Song, Haiyue Wei, Lin Bai, Lei Yang, and Caiyan Jia. Graphalign: Enhancing accurate feature alignment by graph matching for multi-modal 3d object detection. In ICCV, pages 3358–3369, 2023. 2

    2023

  29. [37]

    Corrmatch: Label propagation via correlation matching for semi-supervised semantic segmentation

    Boyuan Sun, Yuqi Yang, Le Zhang, Ming-Ming Cheng, and Qibin Hou. Corrmatch: Label propagation via correlation matching for semi-supervised semantic segmentation. In CVPR, pages 3097–3107, 2024. 3

    2024

  30. [38]

    Spatial context-aware self-attention model for multi-organ segmentation

    Hao Tang, Xingwei Liu, Kun Han, Xiaohui Xie, Xuming Chen, Huang Qian, Yong Liu, Shanlin Sun, and Narisu Bai. Spatial context-aware self-attention model for multi-organ segmentation. In WACV, pages 939–949, 2021. 1

    2021

  31. [39]

    High-resolution 3d abdominal segmentation with random patch network fusion

    Yucheng Tang, Riqiang Gao, Ho Hin Lee, Shizhong Han, Yunqiang Chen, Dashan Gao, Vishwesh Nath, Camilo Bermudez, Michael R Savona, Richard G Abramson, et al. High-resolution 3d abdominal segmentation with random patch network fusion. Medical Image Analysis, 69:101894,

  32. [40]

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

    Antti Tarvainen and Harri Valpola. Mean teachers are better role models: Weight-averaged consistency targets improve semi-supervised deep learning results. NeruIPS, 30, 2017. 3

    2017

  33. [41]

    Advent: Adversarial entropy mini- mization for domain adaptation in semantic segmentation

    Tuan-Hung Vu, Himalaya Jain, Maxime Bucher, Matthieu Cord, and Patrick P´erez. Advent: Adversarial entropy mini- mization for domain adaptation in semantic segmentation. In CVPR, pages 2517–2526, 2019. 7

    2019

  34. [42]

    Towards generic semi- supervised framework for volumetric medical image seg- mentation

    Haonan Wang and Xiaomeng Li. Towards generic semi- supervised framework for volumetric medical image seg- mentation. NeurIPS, 36, 2024. 2

    2024

  35. [43]

    Semi-supervised learning by augmented distribution alignment

    Qin Wang, Wen Li, and Luc Van Gool. Semi-supervised learning by augmented distribution alignment. In ICCV, pages 1466–1475, 2019. 1

    2019

  36. [44]

    Contrastmask: Contrastive learn- ing to segment every thing

    Xuehui Wang, Kai Zhao, Ruixin Zhang, Shouhong Ding, Yan Wang, and Wei Shen. Contrastmask: Contrastive learn- ing to segment every thing. In CVPR, pages 11604–11613,

  37. [45]

    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. In MICCAI, pages 297–306,

  38. [46]

    Mutual consistency learning for semi-supervised medical image segmentation

    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. 2, 3

    2022

  39. [47]

    Exploring smoothness and class-separation for semi-supervised medical image segmentation

    Yicheng Wu, Zhonghua Wu, Qianyi Wu, Zongyuan Ge, and Jianfei Cai. Exploring smoothness and class-separation for semi-supervised medical image segmentation. In MICCAI, pages 34–43, 2022. 1, 6, 7

    2022

  40. [48]

    A global benchmark of algo- rithms for segmenting the left atrium from late gadolinium- enhanced cardiac magnetic resonance imaging

    Zhaohan Xiong, Qing Xia, Zhiqiang Hu, Ning Huang, Cheng Bian, Yefeng Zheng, Sulaiman Vesal, Nishant Ravikumar, Andreas Maier, Xin Yang, et al. A global benchmark of algo- rithms for segmenting the left atrium from late gadolinium- enhanced cardiac magnetic resonance imaging. M...

    2021

  41. [49]

    Spatial tempo- ral graph convolutional networks for skeleton-based action recognition

    Sijie Yan, Yuanjun Xiong, and Dahua Lin. Spatial tempo- ral graph convolutional networks for skeleton-based action recognition. In AAAI, 2018. 3

    2018

  42. [50]

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

    Chenyu You, Weicheng Dai, Yifei Min, Fenglin Liu, David Clifton, S Kevin Zhou, Lawrence Staib, and James Duncan. Rethinking semi-supervised medical image segmentation: A variance-reduction perspective. NeurIPS, 36, 2024. 2

    2024

  43. [51]

    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. InMICCAI, pages 605–613, 2019. 6, 7

    2019

  44. [52]

    Scalable semi-supervised clustering via structural entropy with different constraints

    Guangjie Zeng, Hao Peng, Angsheng Li, Jia Wu, Chunyang Liu, and S Yu Philip. Scalable semi-supervised clustering via structural entropy with different constraints. TKDE, (01):1– 14, 2024. 2

    2024

  45. [53]

    Decoupling the depth and scope of graph neural networks

    Hanqing Zeng, Muhan Zhang, Yinglong Xia, Ajitesh Srivas- tava, Andrey Malevich, Rajgopal Kannan, Viktor Prasanna, Long Jin, and Ren Chen. Decoupling the depth and scope of graph neural networks. NeurIPS, 34:19665–19679, 2021. 3

    2021

  46. [54]

    Affinity attention graph neural network for weakly supervised semantic segmentation

    Bingfeng Zhang, Jimin Xiao, Jianbo Jiao, Yunchao Wei, and Yao Zhao. Affinity attention graph neural network for weakly supervised semantic segmentation. TPAMI, 44(11): 8082–8096, 2021. 2

    2021

  47. [55]

    Deep adversarial net- works for biomedical image segmentation utilizing unanno- tated images

    Yizhe Zhang, Lin Yang, Jianxu Chen, Maridel Fredericksen, David P Hughes, and Danny Z Chen. Deep adversarial net- works for biomedical image segmentation utilizing unanno- tated images. In MICCAI, pages 408–416. Springer, 2017. 7

    2017

  48. [56]

    Graphfpn: Graph feature pyramid network for object detection

    Gangming Zhao, Weifeng Ge, and Yizhou Yu. Graphfpn: Graph feature pyramid network for object detection. In ICCV, pages 2763–2772, 2021. 2

    2021

  49. [57]

    Graph attention guid- ance network with knowledge distillation for semantic seg- mentation of remote sensing images

    Wujie Zhou, Xiaomin Fan, Weiqing Yan, Shengdao Shan, Qiuping Jiang, and Jenq-Neng Hwang. Graph attention guid- ance network with knowledge distillation for semantic seg- mentation of remote sensing images. TGRAS, 2023. 4

    2023

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.