Pith. sign in

REVIEW 4 major objections 5 minor 56 references

This paper claims that medical segmentation noise is spatially concentrated at anatomical boundaries, and that per-pixel meta-learned trust weights combined with a boundary-aware feature-distance loss let segmentation networks stay accurate

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

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

T0 review

2026-08-03 20:35 UTC pith:SJPMCRRE

load-bearing objection Plausible L2B-style meta-reweighting plus a boundary-distance module, but the paper's core weight definition contradicts itself and the proof assumes its conclusion; the experiments are broad but not clean enough to support the headline claims. the 4 major comments →

arxiv 2511.18894 v6 pith:SJPMCRRE submitted 2025-11-24 cs.CV cs.AI

Not All Pixels Are Equal: Pixel-wise Meta-Learning for Medical Segmentation with Noisy Labels

Chenyu Mu , Guihai Chen , Xun Yang , Erkun Yang , Cheng Deng This is my paper
classification cs.CV cs.AI
keywords noisy label learningmedical image segmentationpixel-wise meta-learningboundary uncertaintydynamic center distancelabel noise robustnessbi-level optimizationsegmentation with noisy annotations
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

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

MetaDCSeg claims that medical segmentation noise is not uniform: most annotation errors live at anatomical boundaries, so global noise-robust tricks fail. The paper introduces a two-part framework: a bi-level meta-learning loop that learns a trust weight for every pixel (how much to believe the noisy ground truth vs. the model's own prediction), and a Dynamic Center Distance module that measures each boundary pixel's feature-space distance to foreground, background, and boundary centers and spends extra loss on the ambiguous ones. Across four medical benchmarks with 20–60% synthetic boundary noise, the method consistently outperforms existing baselines, with the largest gains at high noise (e.g., 13.96 mIoU points on PROMISE12). The framework also improves clean-data results, and a 2% clean meta-validation set is enough to drive most of the benefit.

Core claim

The central claim is that the bottleneck in learning from noisy medical labels is spatial heterogeneity: pixels near anatomical boundaries are disproportionately corrupted, and treating all pixels equally—via global confidence thresholds or instance-level reweighting—overfits either noise or background. MetaDCSeg instead meta-learns, for each pixel, a complementary pair of weights α and β that balance the noisy ground-truth label against the model's own pseudo-label, using a small expert-clean validation set to guide the update. Then, using the resulting reliability map γ, it computes feature centers for foreground, background, and boundary regions and defines a composite distance DCD = ‖h−c

What carries the argument

Two coupled mechanisms carry the argument. (1) Pixel-wise bootstrapping loss with meta-learned complementary weights: at each pixel, the training loss is α·CE(p, y_real) + β·CE(p, y_pseudo) with α+β=1, where α (trust in the observed label) and β (trust in the pseudo-label) are optimized in an outer loop against a clean validation set, while the model parameters are optimized in the inner loop. (2) The Dynamic Center Distance (DCD): using γ=α+β as a pixel reliability weight, the method builds feature centers c_fg, c_bg, c_bd for high-confidence, low-confidence, and boundary pixels; each boundary pixel gets DCD = ‖h−c_fg‖·‖h−c_bg‖/(‖h−c_bd‖+ε), and softmax-normalized DCD values weight an extra

Load-bearing premise

The performance claims rest on the assumption that the synthetic corruption pipeline—rotations, morphological erosion/dilation, and ellipse replacement—produces label noise resembling how real annotators actually make boundary errors, and that a small expert-clean meta-validation set is available during training.

What would settle it

Run MetaDCSeg on natural noisy labels, e.g., masks derived from multiple expert annotators with known inter-observer disagreement, without synthetic corruption; if mIoU gains over baselines largely disappear compared to the synthetic-noise experiments, the noise model is the load-bearing element. Conversely, remove the clean meta-set (e.g., replace it with noisy labels) and check whether performance collapses toward baseline; a large drop would show the clean-set requirement is not just an implementation convenience.

Watch this falsifier. Get emailed when new claim-graph text bears on it.

Share X Bluesky LinkedIn Reddit HN

If this is right

  • At 60% synthetic boundary noise, MetaDCSeg retains 86.56% mIoU on MSD Heart and 75.68% on PROMISE12, outperforming prior methods by up to 13.96 mIoU points; the gains grow with noise level.
  • The method does not sacrifice clean data: it exceeds baseline performance on uncorrupted labels (e.g., 93.89% mIoU on MSD Heart), so noise-robust components do not impose a clean-set penalty.
  • Ablations place the meta-learning module as the main driver (removing it costs 4.22 mIoU at 40% noise), with DCD contributing boundary-specific gains and Dice loss enforcing global structural consistency.
  • A clean meta-validation set as small as 2% of the data captures most of the benefit; scaling to 10% adds only 0.28 mIoU while roughly doubling time and memory cost.
  • Inference time is essentially unchanged by the added modules (≈45.2 ms), so the robustness gains do not slow deployment.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • An unstated corollary is that the DCD boundary term is not logically tied to meta-learning: it could be attached to any segmentation loss whose confidence map supplies reliability weights, suggesting a cheaper adapter for existing robust-training pipelines.
  • A testable extension is to inject spatially uniform label noise (random flips of interior pixels) alongside the boundary-targeted protocol; the paper's mechanism predicts a much smaller gain in that regime, since DCD's value depends on noise concentrating at transitions.
  • The boundary center c_bd can be read as a feature-space 'transition centroid', implying DCD measures how strongly a pixel's representation mixes the two classes—an interpretation that could yield uncertainty maps for active learning or clinician QA without new training.
  • Since the online approximation performs a single-step gradient unroll, the method's practical behavior may depend on meta-learning rate and batch composition; replaying the same protocol with different unroll steps (k-step meta-gradients) is a natural robustness probe the paper does not run.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 5 minor

Summary. The paper proposes MetaDCSeg, a medical image segmentation framework designed to be robust to noisy annotations and boundary ambiguity. The method combines a pixel-wise meta-learning scheme that learns per-pixel weights for real and pseudo labels with a Dynamic Center Distance (DCD) module that reweights boundary pixels using distances to foreground, background, and boundary feature centers. The authors report experiments on MSD Heart, MSD Brain, PROMISE12, and Kvasir-SEG under synthetic noise levels of 0%, 20%, 40%, and 60%, and claim consistent improvements over several state-of-the-art baselines, with ablations and complexity analysis. The paper also includes a theoretical convergence analysis in the appendix.

Significance. If the method were precisely specified and the experiments reproducible, the idea of learning spatially varying trust in noisy labels is a reasonable and potentially useful contribution to medical segmentation. The manuscript has several strengths: it reports four datasets, ablates the proposed modules, examines meta-dataset size and computational cost, and provides an anonymous code-link. However, the central reweighting mechanism is described inconsistently across the main text and appendix, the convergence proof is incomplete, and one baseline (MADGNet on PROMISE12) produces implausible results on clean data. These issues are load-bearing for the paper's central claim of consistent, significant state-of-the-art gains, so the paper is not acceptable in its present form.

major comments (4)
  1. [§3.2.3, Eq. (3); §3.3.3, Eq. (9); Appendix A.3, Eqs. (21)–(24)] The central reweighting mechanism is internally inconsistent. §3.2.3 requires α_hw+β_hw=1 with α,β∈[0,1] per pixel, and Eq. (3) is built on that complementary pair. Appendix A.3 updates the two weights independently by gradient descent, rectifies them with max(·,0), and normalizes by Z_i=Σ_{h,w}[α̃+β̃] over the entire image (Eq. 24). After this update, α̂+β̂ is not 1 per pixel and varies spatially. Equation (9) defines γ from the pre-normalization rectified values, while the convergence proof in A.4.1 uses the normalized values. These are three mutually incompatible definitions. Conversely, if α+β=1 were enforced, γ would be identically 1, making the center weighting in Eq. (10) constant and eliminating the DCD contribution. The paper does not state which variant produced Tables 1–4. Because this reweighting is the paper's core contribution, the headline empirical claims are not reproduc
  2. [Appendix A.4.1, Theorem 1] The convergence proof is incomplete and partly circular. Equation (32) replaces a cross term by a sum of squares using 'the fact that meta-weights are updated to align with validation gradients'; no such fact is proven. The one-step update in Eqs. (28)–(30) only shows that α moves in a direction of positive gradient alignment; it does not imply the equality used in Eq. (32), and the β update is not analyzed. Moreover, the claimed condition λ<√(2/(ησ²ML)) does not follow from Eq. (35): λ cancels when the inequality is divided by λ², and the resulting inequality involves sums of inner products with validation gradients that are not bounded by σ and |V| alone. The theorem therefore does not establish monotonic decrease of the validation loss. It should either be proved rigorously or removed/qualified.
  3. [Table 4, PROMISE12] The MADGNet row reports mIoU=8.92 and DSC=16.38 at 0% noise, and mIoU=5.46 at 40% noise. MADGNet is a published state-of-the-art medical segmentation method; near-zero scores on clean expert labels for prostate MRI are implausible and indicate a broken baseline implementation. Its 'collapse' values are therefore not a valid comparison, and the text statements about MADGNet on PROMISE12 are unsupported. The baseline must be rerun and the table corrected. Since the paper claims consistent gains across datasets, a broken baseline in one dataset also casts doubt on the reliability of the other baseline implementations.
  4. [§4.1 and Appendix A.2] All robustness experiments are conducted only under synthetic label corruption (rotation, morphological erosion/dilation, ellipse replacement). The paper defines the target noise as real annotation boundary inconsistency, but no evidence is provided that this protocol captures how human annotators actually err. Without validation on real noisy labels or at least a comparison against annotator disagreement, the claimed transfer of the reported gains to real-world medical annotations is not established. This is not an internal inconsistency, but it is load-bearing for the paper's applied claims.
minor comments (5)
  1. [§4.1 vs. Appendix A.1] The implementation details say 100 epochs total with SGD, while Table 8 lists per-dataset epochs of 120–180 and different learning rates. Please harmonize these statements.
  2. [§4.1] The text mentions a 'GMM threshold β=3.0', but no GMM procedure is defined in the method. Also, α is used both for momentum (α=0.9) and for the meta-learned real-label weight, which is confusing; please use distinct symbols.
  3. [Figure 2] The caption begins with 'Figure 2. Figure 2. Overview...' — remove the duplication.
  4. [A.5.1 vs. A.4.2] Theorem 2 bounds DCD without clipping, but A.5.1 introduces DCD_max clipping; the stability bound should be updated to reflect the clipping mechanism, otherwise the two analyses are not consistent.
  5. [Eq. (14)] The cost efficiency definition is ad hoc and the denominator can be non-positive or ill-defined for small overheads. Please state the assumptions or replace it with standard FLOPs/memory/time reporting.

Circularity Check

0 steps flagged

No significant circularity: the main empirical claim rests on a separate clean meta-validation set and independent test evaluation; self-citations are not load-bearing, and the proof/consistency issues in the appendix are rigor concerns rather than circular reductions.

full rationale

MetaDCSeg's headline result is an empirical comparison on held-out test splits of four benchmarks. The training set is corrupted synthetically, while Dval is explicitly 'held-out samples retaining their original, uncorrupted expert annotations'; the reported metrics are on test data, not on Dval. The pixel-wise weights (α,β) are obtained by bi-level/online optimization against this clean validation set, so the method is not fitting the test outcome and no fitted parameter is later renamed as a prediction. The only author self-citations ([1],[2], both with co-author Erkun Yang) are attached to a general introductory statement and are not load-bearing; no uniqueness theorem or ansatz is imported from them, and the online approximation is explicitly attributed to external work [52]. The internal inconsistency between the per-pixel complementary constraint α+β=1 in §3.2.3 and the global rectification/normalization in Appendix A.3 (Eqs. 23–24) is a reproducibility/definitional defect, not a circular reduction, because the paper does not specify which variant produced Tables 1–4. Likewise, Theorem 1's proof uses an '≈' step ('using the fact that meta-weights are updated to align with validation gradients') that drops the previous-weight term; this is an unsupported approximation in a sketch, not an equality by construction. These issues should be weighed as rigor/validity risks, but they do not make the central empirical claim equivalent to its inputs.

Axiom & Free-Parameter Ledger

6 free parameters · 5 axioms · 0 invented entities

The method depends on the availability of a clean validation set, a well-behaved feature space for boundary centers, and the realism of the synthetic noise protocol. Several hyperparameters are hand-tuned per dataset, and the convergence proof relies on unverified assumptions.

free parameters (6)
  • tau (confidence threshold) = not explicitly given
    Used to decompose foreground/background/boundary regions (Eq. 6-8); value is chosen by hand.
  • tau_dcd (temperature) = 1.0 (default)
    Controls sharpness of softmax in Eq. 12; set empirically.
  • lambda1, lambda2 (loss weights) = per dataset (e.g., lambda_bd, lambda_dice in Table 8)
    Balance boundary refinement and Dice loss in Eq. 13; tuned per dataset.
  • DCD_max = 100
    Clipping threshold in Eq. 46 to stabilize DCD; empirically determined.
  • tau_min = 10 pixels
    Minimum region size for valid center computation in Eq. 47.
  • Per-dataset hyperparameters (Boundary Ratio, Grad Thresh, Scale) = listed in Table 8
    Each dataset has customized settings, implying per-dataset tuning.
axioms (5)
  • domain assumption A small clean validation set (Dval) is available and representative
    The meta-learning weights are optimized using Dval (Sec. 3.1); if such a set is not available in practice, the method cannot be applied.
  • domain assumption Boundary pixels have feature representations that transition between foreground and background centers
    Key insight behind DCD (Sec. 3.3); if features are not discriminative, the distance weighting is meaningless.
  • standard math Bounded gradients and L-Lipschitz smoothness for convergence proof
    Assumptions in Theorem 1 (Appendix A.4.1); not verified for the actual network.
  • domain assumption Pseudo-labels become reliable as training progresses
    The method increasingly trusts pseudo-labels via beta; if pseudo-labels are poorly calibrated, the approach may amplify errors.
  • ad hoc to paper Synthetic noise injection simulates real annotation noise
    Evaluation relies on synthetic corruptions (rotation, morphology, ellipse replacement); if these don't match real-world boundary noise, the reported gains are not transferable.

reviewed 2026-08-03 · how reviews work

0 comments
Cite this review

Pith. "Pith review of Not All Pixels Are Equal: Pixel-wise Meta-Learning for Medical Segmentation with Noisy Labels." pith.science (2026). https://pith.science/paper/SJPMCRRE

@misc{pith2026251118894,
  author       = {Pith},
  title        = {Pith review of: Not All Pixels Are Equal: Pixel-wise Meta-Learning for Medical Segmentation with Noisy Labels},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/SJPMCRRE}},
  note         = {Machine review of arXiv:2511.18894}
}
Share X Bluesky LinkedIn Reddit HN
read the original abstract

Medical image segmentation is crucial for clinical applications, but it is frequently disrupted by noisy annotations and ambiguous anatomical boundaries, limiting its application in real-world scenarios. Existing methods often directly adapt noisy label learning techniques designed for instance classification, overlooking the pixel-wise heterogeneity in medical segmentation with its spatially and anatomically varying difficulties. Consequently, global assumptions or simple confidence metrics fail to address these local variations, leaving boundary ambiguities unresolved. To address this issue, we propose MetaDCSeg, a robust framework that dynamically learns optimal pixel-wise weights to suppress the influence of noisy labels while preserving reliable annotations. By explicitly modeling boundary uncertainty through a Dynamic Center Distance (DCD) mechanism, our approach utilizes weighted feature distances for foreground, background, and boundary centers, directing the model's attention toward hard-to-segment pixels near ambiguous boundaries. This strategy enables more precise handling of structural boundaries, which are often overlooked by existing methods, and significantly enhances segmentation performance. Extensive experiments across four benchmark datasets with varying noise levels demonstrate that MetaDCSeg outperforms existing state-of-the-art methods.

Figures

Figures reproduced from arXiv: 2511.18894 by Cheng Deng, Chenyu Mu, Erkun Yang, Guihai Chen, Xun Yang.

Figure 1
Figure 1. Figure 1: Boundary challenges in medical image segmenta [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Figure 2. Overview of the MetaDCSeg architecture. The Meta-Learning Module dynamically generates [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Spatial distribution and distance-dependency of [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Visual comparison of segmentation methods on Kvasir-SEG, PROMISE12, MSD Heart, and MSD Brain [PITH_FULL_IMAGE:figures/full_fig_p006_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: 5-Fold Cross-Validation (MSD Heart, 40% Noise) [PITH_FULL_IMAGE:figures/full_fig_p008_5.png] view at source ↗

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

56 extracted references · 3 linked inside Pith

  1. [1]

    A parametrical model for instance-dependent label noise.IEEE Transactions on Pattern Analysis and Machine Intelligence, 45(12):14055–14068, 2023

    Shuo Yang, Songhua Wu, Erkun Yang, Bo Han, Yang Liu, Min Xu, Gang Niu, and Tongliang Liu. A parametrical model for instance-dependent label noise.IEEE Transactions on Pattern Analysis and Machine Intelligence, 45(12):14055–14068, 2023

    2023

  2. [2]

    Subclass-dominant label noise: a counterexample for the success of early stopping.Advances in Neural Information Processing Systems, 36:68343–68366, 2023

    Yingbin Bai, Zhongyi Han, Erkun Yang, Jun Yu, Bo Han, Dadong Wang, and Tongliang Liu. Subclass-dominant label noise: a counterexample for the success of early stopping.Advances in Neural Information Processing Systems, 36:68343–68366, 2023. 9 Running Title for Header

    2023

  3. [3]

    Image segmentation technology based on transformer in medical decision- making system.IET Image Processing, 17(10):3040–3054, 2023

    Keke He, Fangfang Gou, and Jia Wu. Image segmentation technology based on transformer in medical decision- making system.IET Image Processing, 17(10):3040–3054, 2023

    2023

  4. [4]

    U-net: Convolutional networks for biomedical image segmentation

    Olaf Ronneberger, Philipp Fischer, and Thomas Brox. U-net: Convolutional networks for biomedical image segmentation. InInternational Conference on Medical image computing and computer-assisted intervention, pages 234–241. Springer, 2015

    2015

  5. [5]

    Unet++: A nested u-net architecture for medical image segmentation

    Zongwei Zhou, Md Mahfuzur Rahman Siddiquee, Nima Tajbakhsh, and Jianming Liang. Unet++: A nested u-net architecture for medical image segmentation. InDeep Learning in Medical Image Analysis and Multimodal Learning for Clinical Decision Support, pages 3–11. Springer, 2018

    2018

  6. [6]

    nnu-net: a self-configuring method for deep learning-based biomedical image segmentation.Nature methods, 18(2):203–211, 2021

    Fabian Isensee, Paul F Jaeger, Simon AA Kohl, Jens Petersen, and Klaus H Maier-Hein. nnu-net: a self-configuring method for deep learning-based biomedical image segmentation.Nature methods, 18(2):203–211, 2021

    2021

  7. [7]

    The HAM10000 dataset, a large collection of multi-source dermatoscopic images of common pigmented skin lesions.Scientific Data, 5:180161, 2018

    Philipp Tschandl, Cliff Rosendahl, and Harald Kittler. The HAM10000 dataset, a large collection of multi-source dermatoscopic images of common pigmented skin lesions.Scientific Data, 5:180161, 2018

    2018

  8. [8]

    The multimodal brain tumor image segmentation benchmark (brats).IEEE Transactions on Medical Imaging, 34(10):1993–2024, 2015

    Bjoern H Menze, Andras Jakab, Stefan Bauer, Jayashree Kalpathy-Cramer, Keyvan Farahani, Justin Kirby, Yuliya Burren, Nils Porz, Johannes Slotboom, Roland Wiest, et al. The multimodal brain tumor image segmentation benchmark (brats).IEEE Transactions on Medical Imaging, 34(10):1993–2024, 2015

    1993

  9. [9]

    The liver tumor segmentation benchmark (lits)

    Patrick Bilic, Patrick F Christ, Eugene V orontsov, Grzegorz Chlebus, Hao Chen, Qi Dou, Chi-Wing Fu, Xiahai Han, Pheng-Ann Heng, Jürgen Hesser, et al. The liver tumor segmentation benchmark (lits). InInternational Conference on Medical Image Computing and Computer-Assisted Intervention, pages 980–989. Springer, 2019

    2019

  10. [10]

    Evaluation of algorithms for multi-modality whole heart segmentation: an open-access grand challenge.Medical Image Analysis, 70:101937, 2021

    Xiahai Zhuang, Li Li, Christian Payer, Orjan Smedby, Mattias P Heinrich, Martin Urschler, Guang Yang, Pheng-Ann Heng, Shuo Li, Alejandro F Frangi, et al. Evaluation of algorithms for multi-modality whole heart segmentation: an open-access grand challenge.Medical Image Analysis, 70:101937, 2021

    2021

  11. [11]

    Illumination-based transformations improve skin lesion segmentation in dermoscopic images

    Kumar Abhishek, Ghassan Hamarneh, and Mark S Drew. Illumination-based transformations improve skin lesion segmentation in dermoscopic images. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition workshops, pages 728–729, 2020

    2020

  12. [12]

    Medical image segmentation: A comprehensive review of deep learning-based methods.Tomography, 11(5):52, 2025

    Yuxiao Gao, Yang Jiang, Yanhong Peng, Fujiang Yuan, Xinyue Zhang, and Jianfeng Wang. Medical image segmentation: A comprehensive review of deep learning-based methods.Tomography, 11(5):52, 2025

    2025

  13. [13]

    Augmenting medical imaging: A comprehensive catalogue of 65 techniques for enhanced data analysis.arXiv preprint arXiv:2303.01178, 2023

    Manuel Cossio. Augmenting medical imaging: A comprehensive catalogue of 65 techniques for enhanced data analysis.arXiv preprint arXiv:2303.01178, 2023

    Pith/arXiv arXiv 2023

  14. [14]

    A survey on image data augmentation for deep learning.Journal of Big Data, 6(1):60, 2019

    Connor Shorten and Taghi M Khoshgoftaar. A survey on image data augmentation for deep learning.Journal of Big Data, 6(1):60, 2019

    2019

  15. [15]

    Directional connectivity-based segmentation of medical images

    Ziyun Yang and Sina Farsiu. Directional connectivity-based segmentation of medical images. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 11525–11535, 2023

    2023

  16. [16]

    Maxformer: Enhanced transformer for medical image segmentation with multi-attention and multi-scale features fusion.Knowledge-Based Systems, 280:110987, 2023

    Zhiwei Liang, Kui Zhao, Gang Liang, Siyu Li, Yifei Wu, and Yiping Zhou. Maxformer: Enhanced transformer for medical image segmentation with multi-attention and multi-scale features fusion.Knowledge-Based Systems, 280:110987, 2023

    2023

  17. [17]

    Translation consistent semi-supervised segmentation for 3d medical images.IEEE Transactions on Medical Imaging, 2024

    Yuyuan Liu, Yu Tian, Chong Wang, Yuanhong Chen, Fengbei Liu, Vasileios Belagiannis, and Gustavo Carneiro. Translation consistent semi-supervised segmentation for 3d medical images.IEEE Transactions on Medical Imaging, 2024

    2024

  18. [18]

    Modality-agnostic domain generalizable medical image segmentation by multi-frequency in multi-scale attention

    Ju-Hyeon Nam, Nur Suriza Syazwany, Su Jung Kim, and Sang-Chul Lee. Modality-agnostic domain generalizable medical image segmentation by multi-frequency in multi-scale attention. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 11480–11491, 2024

    2024

  19. [19]

    Condseg: A general medical image segmentation framework via contrast-driven feature enhancement

    Mengqi Lei, Haochen Wu, Xinhua Lv, and Xin Wang. Condseg: A general medical image segmentation framework via contrast-driven feature enhancement. InProceedings of the AAAI Conference on Artificial Intelligence, number 5, pages 4571–4579, 2025

    2025

  20. [20]

    Eshmam Rayed, S

    Md. Eshmam Rayed, S. M. Sajibul Islam, Sadia Islam Niha, Jamin Rahman Jim, Md Mohsin Kabir, and M. F. Mridha. Deep learning for medical image segmentation: State-of-the-art advancements and challenges.Informatics in Medicine Unlocked, 47:101504, 2024

    2024

  21. [21]

    Medical image segmentation using deep learning: A survey.IET Image Processing, 16(5):1243–1267, 2022

    Risheng Wang, Tao Lei, Ruixia Cui, Bingtao Zhang, Hongying Meng, and Asoke K Nandi. Medical image segmentation using deep learning: A survey.IET Image Processing, 16(5):1243–1267, 2022

    2022

  22. [22]

    Medical image segmentation using deep semantic-based methods: A review of techniques, applications and emerging trends.Information Fusion, 90:316–352, 2023

    Imran Qureshi, Junhua Yan, Qaisar Abbas, Kashif Shaheed, Awais Bin Riaz, Abdul Wahid, Muhammad Waseem Jan Khan, and Piotr Szczuko. Medical image segmentation using deep semantic-based methods: A review of techniques, applications and emerging trends.Information Fusion, 90:316–352, 2023. 10 Running Title for Header

    2023

  23. [23]

    Swin SMT: Global sequential modeling for enhancing 3d medical image segmentation

    Szymon Płotka, Maciej Chrabaszcz, and Przemyslaw Biecek. Swin SMT: Global sequential modeling for enhancing 3d medical image segmentation. InMedical Image Computing and Computer-Assisted Intervention (MICCAI 2024), LNCS 15008, pages 689–699. Springer, 2024

    2024

  24. [24]

    Medical image segmentation by combining feature enhancement swin transformer and upernet.Scientific Reports, 15(14565), 2025

    Lin Zhang, Xiaochun Yin, Xuqi Liu, and Zengguang Liu. Medical image segmentation by combining feature enhancement swin transformer and upernet.Scientific Reports, 15(14565), 2025

    2025

  25. [25]

    ProMISe: Prompt-driven 3d medical image segmentation using pretrained image foundation models.arXiv preprint, 2023

    Hao Li, Han Liu, Dewei Hu, Jiacheng Wang, and Ipek Oguz. ProMISe: Prompt-driven 3d medical image segmentation using pretrained image foundation models.arXiv preprint, 2023

    2023

  26. [26]

    Medical SAM adapter: Adapting segment anything model for medical image segmentation.Medical Image Analysis, 102:103547, 2025

    Junde Wu, Ziyue Wang, Mingxuan Hong, Wei Ji, Yuanpei Liu, Huazhu Fu, Yanwu Xu, and Yueming Jin. Medical SAM adapter: Adapting segment anything model for medical image segmentation.Medical Image Analysis, 102:103547, 2025

    2025

  27. [27]

    Latent diffusion for medical image segmentation: End-to-end learning for fast sampling and accuracy.arXiv preprint, 2025

    Fahim Ahmed Zaman, Mathews Jacob, Amanda Chang, Kan Liu, Milan Sonka, and Xiaodong Wu. Latent diffusion for medical image segmentation: End-to-end learning for fast sampling and accuracy.arXiv preprint, 2025

    2025

  28. [28]

    Conditional diffusion model with spatial attention and latent embedding for medical image segmentation.arXiv preprint, 2025

    Behzad Hejrati, Soumyanil Banerjee, Carri Glide-Hurst, and Ming Dong. Conditional diffusion model with spatial attention and latent embedding for medical image segmentation.arXiv preprint, 2025

    2025

  29. [29]

    Le Zhang, Fuping Wu, Kevin Bronik, and Bartlomiej W. Papiez. DiffuSeg: Domain-driven diffusion for medical image segmentation.IEEE Journal of Biomedical and Health Informatics, 29(5):3619–3631, 2025

    2025

  30. [30]

    Exploring the robustness of in-context learning with noisy labels

    Chen Cheng, Xinzhi Yu, Haodong Wen, Jingsong Sun, Guanzhang Yue, Yihao Zhang, and Zeming Wei. Exploring the robustness of in-context learning with noisy labels. InICASSP 2025-2025 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pages 1–5. IEEE, 2025

    2025

  31. [31]

    Estimating instance-dependent label-noise transition matrix using a deep neural network

    Jialu Wang, Eric Xin Wang, and Yang Liu. Estimating instance-dependent label-noise transition matrix using a deep neural network. InICML, 2022

    2022

  32. [32]

    Distilling effective supervision from severe label noise

    Zizhao Zhang, Han Zhang, Sercan O Arik, Honglak Lee, and Tomas Pfister. Distilling effective supervision from severe label noise. InCVPR, pages 9294–9303, 2020

    2020

  33. [33]

    Fnbench: Benchmarking robust federated learning against noisy labels.arXiv preprint arXiv:2505.06684, 2025

    Xuefeng Jiang, Jia Li, Nannan Wu, Zhiyuan Wu, Xujing Li, Sheng Sun, Gang Xu, Yuwei Wang, Qi Li, and Min Liu. Fnbench: Benchmarking robust federated learning against noisy labels.arXiv preprint arXiv:2505.06684, 2025

    Pith/arXiv arXiv 2025

  34. [34]

    Dividemix: Learning with noisy labels as semi-supervised learning.ICLR, 2020

    Junnan Li, Richard Socher, and Steven CH Hoi. Dividemix: Learning with noisy labels as semi-supervised learning.ICLR, 2020

    2020

  35. [35]

    Unicon: Combating label noise through uniform selection and contrastive learning

    Nazmul Karim, Mamshad Nayeem Rizve, Nazanin Rahnavard, Ajmal Mian, and Mubarak Shah. Unicon: Combating label noise through uniform selection and contrastive learning. InCVPR, pages 9676–9686, 2022

    2022

  36. [36]

    Contrast to divide: Self-supervised pre-training for learning with noisy labels

    Evgenii Zheltonozhskii, Chaim Baskin, Avi Mendelson, Alex M Bronstein, and Or Litany. Contrast to divide: Self-supervised pre-training for learning with noisy labels. InWACV, pages 1657–1667, 2022

    2022

  37. [37]

    Adaptive integration of partial label learning and negative learning for enhanced noisy label learning

    Mengmeng Sheng, Zeren Sun, Zhenhuang Cai, Tao Chen, Yichao Zhou, and Yazhou Yao. Adaptive integration of partial label learning and negative learning for enhanced noisy label learning. InProceedings of the AAAI conference on artificial intelligence, volume 38, pages 4820–4828, 2024

    2024

  38. [38]

    Learning to learn from noisy labeled data

    Junnan Li, Yongkang Wong, Qi Zhao, and Mohan S Kankanhalli. Learning to learn from noisy labeled data. In CVPR, pages 5051–5059, 2019

    2019

  39. [39]

    Advances and challenges in meta-learning: A technical review.IEEE transactions on pattern analysis and machine intelligence, 46(7):4763–4779, 2024

    Anna Vettoruzzo, Mohamed-Rafik Bouguelia, Joaquin Vanschoren, Thorsteinn Rögnvaldsson, and KC Santosh. Advances and challenges in meta-learning: A technical review.IEEE transactions on pattern analysis and machine intelligence, 46(7):4763–4779, 2024

    2024

  40. [40]

    Learning to purify noisy labels via meta soft label corrector

    Yichen Wu, Jun Shu, Qi Xie, Qian Zhao, and Deyu Meng. Learning to purify noisy labels via meta soft label corrector. InAAAI, number 12, pages 10388–10396, 2021

    2021

  41. [41]

    Hieu Pham, Mingxing Guan, Zhen Liu, and Quoc V . Le. Meta-learning with warped gradient descent for noisy supervision. InProceedings of the International Conference on Learning Representations (ICLR), 2019

    2019

  42. [42]

    Learning to balance: Meta-learning for imbalanced and noisy medical image segmentation

    Yuhao Zhang, Guotai Yang, Liang Xu, Yong Xu, and Yefeng Zheng. Learning to balance: Meta-learning for imbalanced and noisy medical image segmentation. InInternational Conference on Medical Image Computing and Computer-Assisted Intervention (MICCAI), pages 276–286. Springer, 2021

    2021

  43. [43]

    Joint meta-learning for robust medical image segmentation with noisy labels

    Yanyun Guo, Xindi Liu, Liansheng Wang, and Qian Li. Joint meta-learning for robust medical image segmentation with noisy labels. InProceedings of the AAAI Conference on Artificial Intelligence, number 1, pages 132–140, 2022. 11 Running Title for Header

    2022

  44. [44]

    Mgl: Memory-guided learner for robust medical image segmentation with noisy labels.Medical Image Analysis, 85:102748, 2023

    Yahui Liu, Qiang Sun, Mingzhe Li, Shuo Wang, and Shaoting Zhang. Mgl: Memory-guided learner for robust medical image segmentation with noisy labels.Medical Image Analysis, 85:102748, 2023

    2023

  45. [45]

    Cmw-net: Learning a class-aware sample weighting mapping for robust deep learning.IEEE Trans

    Jun Shu, Xiang Yuan, Deyu Meng, and Zongben Xu. Cmw-net: Learning a class-aware sample weighting mapping for robust deep learning.IEEE Trans. Pattern Anal. Mach. Intell., 45(10):11521–11539, 2023

    2023

  46. [46]

    Learning from noisy labels with decoupled meta label purifier

    Yuanpeng Tu, Boshen Zhang, Yuxi Li, Liang Liu, Jian Li, Yabiao Wang, Chengjie Wang, and Cai Rong Zhao. Learning from noisy labels with decoupled meta label purifier. InCVPR, pages 19934–19943, 2023

    2023

  47. [47]

    Meta label correction for noisy label learning

    Guoqing Zheng, Ahmed Hassan Awadallah, and Susan Dumais. Meta label correction for noisy label learning. In AAAI, number 12, pages 11053–11061, 2021

    2021

  48. [48]

    Meta-learning collaborative optimization for lifetime prediction of lithium-ion batteries considering label noise.Journal of Energy Storage, 107:114928, 2025

    Guisong Wang, Cong Wang, Yunxia Chen, and Jie Liu. Meta-learning collaborative optimization for lifetime prediction of lithium-ion batteries considering label noise.Journal of Energy Storage, 107:114928, 2025

    2025

  49. [49]

    Meta-learning with elastic prototypical network for fault transfer diagnosis of bearings under unstable speeds.Reliability Engineering & System Safety, 245:110001, 2024

    Jingjie Luo, Haidong Shao, Jian Lin, and Bin Liu. Meta-learning with elastic prototypical network for fault transfer diagnosis of bearings under unstable speeds.Reliability Engineering & System Safety, 245:110001, 2024

    2024

  50. [50]

    Application of an electronic tongue combined with meta-learning for rice origin detection.Engineering Applications of Artificial Intelligence, 156:111174, 2025

    Litong Chen, Zihan Wang, Zhiqiang Wang, Xuerui Bai, Jingyu Ma, Chuanzheng Liu, Tao Sun, and Yubin Lan. Application of an electronic tongue combined with meta-learning for rice origin detection.Engineering Applications of Artificial Intelligence, 156:111174, 2025

    2025

  51. [51]

    Learning to reweight examples for robust deep learning

    Mengye Ren, Wenyuan Zeng, Bin Yang, and Raquel Urtasun. Learning to reweight examples for robust deep learning. InInternational conference on machine learning, pages 4334–4343. PMLR, 2018

    2018

  52. [52]

    L2b: Learning to bootstrap robust models for combating label noise

    Yuyin Zhou, Xianhang Li, Fengze Liu, Qingyue Wei, Xuxi Chen, Lequan Yu, Cihang Xie, Matthew P Lungren, and Lei Xing. L2b: Learning to bootstrap robust models for combating label noise. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 23523–23533, 2024

    2024

  53. [53]

    Pixel difference networks for efficient edge detection

    Zhuo Su, Wenzhe Liu, Zitong Yu, Dewen Hu, Qing Liao, Qi Tian, Matti Pietikäinen, and Li Liu. Pixel difference networks for efficient edge detection. InProceedings of the IEEE/CVF international conference on computer vision, pages 5117–5127, 2021

    2021

  54. [54]

    Simpson, Michela Antonelli, Spyridon Bakas, Michel Bilello, Keyvan Farahani, Bram van Ginneken, Annette Kopp–Schneider, Bennett A

    Amber L. Simpson, Michela Antonelli, Spyridon Bakas, Michel Bilello, Keyvan Farahani, Bram van Ginneken, Annette Kopp–Schneider, Bennett A. Landman, Geert Litjens, Bjoern H. Menze, Olaf Ronneberger, Ronald M. Summers, Patrick Bilic, Patrick F. Christ, Richard K. G. Do, Marc Gollub, Jennifer Golia-Pernicka, Stephan H. Heckers, William R. Jarnagin, Maureen ...

    Pith/arXiv arXiv 1902

  55. [55]

    Evaluation of prostate segmentation algorithms for MRI: The PROMISE12 challenge.Medical Image Analysis, 18(2):359–373, 2014

    Geert Litjens, Robert Tóth, Wendy van de Ven, Caroline Hoeks, Sjoerd Kerkstra, Bram van Ginneken, and Henkjan Huisman. Evaluation of prostate segmentation algorithms for MRI: The PROMISE12 challenge.Medical Image Analysis, 18(2):359–373, 2014

    2014

  56. [56]

    Smedsrud, Michael A

    Debesh Jha, Pia H. Smedsrud, Michael A. Riegler, Pål Halvorsen, Thomas de Lange, Dag Johansen, and Håvard D. Johansen. Kvasir-seg: A segmented polyp dataset. InMultiMedia Modeling: 26th International Conference, MMM 2020, Daejeon, South Korea, January 5–8, 2020, Proceedings, Part II, volume 11962 ofLecture Notes in Computer Science, pages 451–462. Springe...

    2020

This paper was first reviewed by deepseek-v4-flash on August 3, 2026.