mlsanityData to 7 Oct 2026

ZScribbleSeg: A comprehensive segmentation framework with modeling of efficient annotation and maximization of scribble supervision

arXiv 2605.06266 ↗First posted 7 May 2026cs.CVcited by 0

Not on the map. Why?

Abstract (arXiv)

Curating fully annotated datasets for medical image segmentation is labour-intensive and expertise-demanding. To alleviate this problem, prior studies have explored scribble annotations for weakly supervised segmentation. Existing solutions mainly compute losses on annotated areas and generate pseudo labels by propagating annotations to adjacent regions. However, these methods often suffer from inaccurate and unrealistic segmentations due to insufficient supervision and incomplete shape information. In contrast, we first investigate the principle of good scribble annotations, which leads to efficient scribble forms via supervision maximization and randomness simulation. We further introduce regularization terms to encode the spatial relationship and the shape constraints, where the EM algorithm is utilized to estimate the mixture ratios of label classes. These ratios are critical in identifying the unlabeled pixels for each class and correcting erroneous predictions, thus the accurate estimation lays the foundation for the incorporation of spatial prior. Finally, we integrate the efficient scribble supervision with the prior into a framework, referred to as ZScribbleSeg, and apply it to multiple scenarios. Leveraging only scribble annotations, ZScribbleSeg achieves competitive performance on six segmentation tasks including ACDC, MSCMRseg, BTCV, MyoPS, Decathlon-BrainTumor and Decathlon-Prostate. Our code will be released via https://github.com/DLwbm123/ZScribbleSeg.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Builds on 39 of its 70 references that are in the map. First posted on arXiv 7 May 2026.

Builds on

39 of 70 references.

References · 39

  1. 2017mixup: Beyond Empirical Risk Minimizationcited by 5,155
  2. 2015Learning Deep Features for Discriminative Localizationcited by 2,430
  3. 2021The Medical Segmentation Decathloncited by 620
  4. 2020Semi-Supervised Semantic Segmentation with Cross-Consistency Trainingcited by 375
  5. 2018A Probabilistic U-Net for Segmentation of Ambiguous Imagescited by 281
  6. 2020Puzzle Mix: Exploiting Saliency and Local Statistics for Optimal Mixupcited by 260
  7. 2019Embracing Imperfect Datasets: A Review of Deep Learning Solutions for Medical Image Segmentationcited by 258
  8. 2019Understanding Deep Networks via Extremal Perturbations and Smooth Maskscited by 257
  9. 2016Simple Does It: Weakly Supervised Instance and Semantic Segmentationcited by 226
  10. 2017Positive-Unlabeled Learning with Non-Negative Risk Estimatorcited by 217
  11. 2016Multivariate mixture model for myocardium segmentation combining multi-source imagescited by 133
  12. 2021Co-Mixup: Saliency Guided Joint Mixup with Supermodular Diversitycited by 127
27 more
  1. 2018On Regularized Losses for Weakly-supervised CNN Segmentationcited by 121
  2. 2016Mixture Proportion Estimation via Kernel Embedding of Distributionscited by 67
  3. 2022CycleMix: A Holistic Strategy for Medical Image Segmentation from Scribble Supervisioncited by 64
  4. 2022Scribble-Supervised Medical Image Segmentation via Dual-Branch Network and Dynamically Mixed Pseudo Labels Supervisioncited by 64
  5. 2018Constrained-CNN losses for weakly supervised segmentationcited by 62
  6. 2020FocalMix: Semi-Supervised Learning for 3D Medical Image Detectioncited by 52
  7. 2021Mixture Proportion Estimation and PU Learning: A Modern Approachcited by 52
  8. 2017Constrained Deep Weak Supervision for Histopathology Image Segmentationcited by 51
  9. 2018Learning to Segment Medical Images with Scribble-Supervision Alonecited by 48
  10. 2021Affinity Attention Graph Neural Network for Weakly Supervised Semantic Segmentationcited by 47
  11. 2019Semi-Supervised and Task-Driven Data Augmentationcited by 40
  12. 2019Gated CRF Loss for Weakly Supervised Semantic Image Segmentationcited by 40
  13. 2020Going to Extremes: Weakly Supervised Medical Image Segmentationcited by 38
  14. 2017An Exploration of 2D and 3D Deep Learning Techniques for Cardiac MR Image Segmentationcited by 36
  15. 2019Scribble-based Hierarchical Weakly Supervised Learning for Brain Tumor Segmentationcited by 32
  16. 2019Cardiac Segmentation from LGE MRI Using Deep Neural Network Incorporating Shape and Spatial Priorscited by 27
  17. 2020Deep Interpretable Classification and Weakly-Supervised Segmentation of Histology Images via Max-Min Uncertaintycited by 26
  18. 2022ShapePU: A New PU Learning Framework Regularized by Global Consistency for Scribble Supervised Cardiac Segmentationcited by 26
  19. 2024ScribFormer: Transformer Makes CNN Work Better for Scribble-based Medical Image Segmentationcited by 24
  20. 2023WeakPolyp: You Only Look Bounding Box for Polyp Segmentationcited by 21
  21. 2022Nuclei Segmentation with Point Annotations from Pathology Images via Self-Supervised Learning and Co-Trainingcited by 16
  22. 2020ACCL: Adversarial constrained-CNN loss for weakly supervised medical image segmentationcited by 13
  23. 2024SUMix: Mixup with Semantic and Uncertain Informationcited by 12
  24. 2022ECONet: Efficient Convolutional Online Likelihood Network for Scribble-based Interactive Segmentationcited by 11
  25. 2017Employing Weak Annotations for Medical Image Analysis Problemscited by 7
  26. 2024FedLPPA: Learning Personalized Prompt and Aggregation for Federated Weakly-supervised Medical Image Segmentationcited by 6
  27. 2023Volumetric Medical Image Segmentation via Scribble Annotations and Shape Priorscited by 1

Built on by

Not cited yet.