mlsanityData to 13 Oct 2026

Generalised Dice overlap as a deep learning loss function for highly unbalanced segmentations

arXiv 1707.03237 ↗First posted 11 Jul 2017cs.CVcited by 615

Not on the map. Why?

Abstract (arXiv)

Deep-learning has proved in recent years to be a powerful tool for image analysis and is now widely used to segment both 2D and 3D medical images. Deep-learning segmentation frameworks rely not only on the choice of network architecture but also on the choice of loss function. When the segmentation process targets rare observations, a severe class imbalance is likely to occur between candidate labels, thus resulting in sub-optimal performance. In order to mitigate this issue, strategies such as the weighted cross-entropy function, the sensitivity function or the Dice loss function, have been proposed. In this work, we investigate the behavior of these loss functions and their sensitivity to learning rate tuning in the presence of different rates of label imbalance across 2D and 3D segmentation tasks. We also propose to use the class re-balancing properties of the Generalized Dice overlap, a known metric for segmentation assessment, as a robust and accurate deep-learning loss function for unbalanced tasks.

Key facts. Cited by 615 papers in the map (data through 13 Oct 2026). First posted on arXiv 11 Jul 2017.

Builds on

No reference list yet.

Built on by

Cited by year (2026 to 13 Oct)

Citing papers per year: 2021: 83; 2022: 81; 2023: 116; 2024: 127; 2025: 121; 2026: 87.

615 papers.

Sort:

By area:

  1. 2021Common Objects in 3D: Large-Scale Learning and Evaluation of Real-life 3D Category Reconstructioncited by 648
  2. 2023Ego-Exo4D: Understanding Skilled Human Activity from First- and Third-Person Perspectivescited by 482
  3. 2022MONAI: An open-source framework for deep learning in healthcarecited by 443
  4. 2022Medical Image Segmentation Review: The success of U-Netcited by 191
  5. 2021UNetFormer: A UNet-like Transformer for Efficient Semantic Segmentation of Remote Sensing Urban Scene Imagerycited by 166
  6. 2024FakeShield: Explainable Image Forgery Detection and Localization via Multi-modal Large Language Modelscited by 112
  7. 2021Recent advances and clinical applications of deep learning in medical image analysiscited by 104
  8. 2024PSALM: Pixelwise SegmentAtion with Large Multi-Modal Modelcited by 100
  9. 2023UniverSeg: Universal Medical Image Segmentationcited by 91
  10. 2021Unified Focal loss: Generalising Dice and cross entropy-based losses to handle class imbalanced medical image segmentationcited by 88
  11. 2022Federated Learning Enables Big Data for Rare Cancer Boundary Detectioncited by 83
  12. 2024Infrared Small Target Detection with Scale and Location Sensitivitycited by 73
38 more
  1. 2024MiM-ISTD: Mamba-in-Mamba for Efficient Infrared Small Target Detectioncited by 60
  2. 2023PerceptionGPT: Effectively Fusing Visual Perception into LLMcited by 59
  3. 2021Common Limitations of Image Processing Metrics: A Picture Storycited by 58
  4. 2023Online Detection of AI-Generated Imagescited by 55
  5. 2022ImageCAS: A Large-Scale Dataset and Benchmark for Coronary Artery Segmentation based on Computed Tomography Angiography Imagescited by 54
  6. 2023SAR-RARP50: Segmentation of surgical instrumentation and Action Recognition on Robot-Assisted Radical Prostatectomy Challengecited by 54
  7. 2023Hierarchical Open-vocabulary Universal Image Segmentationcited by 53
  8. 2023ISBNet: a 3D Point Cloud Instance Segmentation Network with Instance-aware Sampling and Box-aware Dynamic Convolutioncited by 52
  9. 2023Deep Learning Approaches for Data Augmentation in Medical Imaging: A Reviewcited by 48
  10. 2024CDMamba: Incorporating Local Clues into Mamba for Remote Sensing Image Binary Change Detectioncited by 48
  11. 2024Make It Count: Text-to-Image Generation with an Accurate Number of Objectscited by 47
  12. 2024SplineGS: Robust Motion-Adaptive Spline for Real-Time Dynamic 3D Gaussians from Monocular Videocited by 44
  13. 2021PolyWorld: Polygonal Building Extraction with Graph Neural Networks in Satellite Imagescited by 43
  14. 2021Active Boundary Loss for Semantic Segmentationcited by 40
  15. 2022TraSeTR: Track-to-Segment Transformer with Contrastive Query for Instance-level Instrument Segmentation in Robotic Surgerycited by 40
  16. 2023EdgeSAM: Prompt-In-the-Loop Distillation for SAMcited by 40
  17. 2023DreamTeacher: Pretraining Image Backbones with Deep Generative Modelscited by 37
  18. 2024RobustSAM: Segment Anything Robustly on Degraded Imagescited by 36
  19. 2024FaceXFormer: A Unified Transformer for Facial Analysiscited by 34
  20. 2024Pinwheel-shaped Convolution and Scale-based Dynamic Loss for Infrared Small Target Detectioncited by 34
  21. 2021Enhanced U-Net: A Feature Enhancement Network for Polyp Segmentationcited by 33
  22. 2023UnScene3D: Unsupervised 3D Instance Segmentation for Indoor Scenescited by 31
  23. 2023Directional Connectivity-based Segmentation of Medical Imagescited by 30
  24. 2023OCELOT: Overlapped Cell on Tissue Dataset for Histopathologycited by 29
  25. 2023BiomedJourney: Counterfactual Biomedical Image Generation by Instruction-Learning from Multimodal Patient Journeyscited by 29
  26. 2023Open-vocabulary object 6D pose estimationcited by 29
  27. 2021Deep Learning methods for automatic evaluation of delayed enhancement-MRI. The results of the EMIDEC challengecited by 28
  28. 2023DINAR: Diffusion Inpainting of Neural Textures for One-Shot Human Avatarscited by 28
  29. 2025Scaling Vision Pre-Training to 4K Resolutioncited by 27
  30. 2021Are we using appropriate segmentation metrics? Identifying correlates of human expert perception for CNN training beyond rolling the DICE coefficientcited by 26
  31. 2024SUGAR: Pre-training 3D Visual Representations for Roboticscited by 26
  32. 2021GaNDLF: A Generally Nuanced Deep Learning Framework for Scalable End-to-End Clinical Workflows in Medical Imagingcited by 25
  33. 2021Two-Stage Mesh Deep Learning for Automated Tooth Segmentation and Landmark Localization on 3D Intraoral Scanscited by 24
  34. 2021Building extraction with vision transformercited by 22
  35. 2022MyoPS-Net: Myocardial Pathology Segmentation with Flexible Combination of Multi-Sequence CMR Imagescited by 22
  36. 2025EdgeTAM: On-Device Track Anything Modelcited by 21
  37. 2025PanSt3R: Multi-view Consistent Panoptic Segmentationcited by 20
  38. 2021Striking the Right Balance: Recall Loss for Semantic Segmentationcited by 19

50 most cited of 615 All 615 (JSON)