mlsanityData to 7 Oct 2026

Inter Extreme Points Geodesics for End-to-End Weakly Supervised Image Segmentation

arXiv 2107.00583 ↗First posted 1 Jul 2021cs.CVcited by 21

Not on the map. Why?

Abstract (arXiv)

We introduce textitInExtremIS, a weakly supervised 3D approach to train a deep image segmentation network using particularly weak train-time annotations: only 6 extreme clicks at the boundary of the objects of interest. Our fully-automatic method is trained end-to-end and does not require any test-time annotations. From the extreme points, 3D bounding boxes are extracted around objects of interest. Then, deep geodesics connecting extreme points are generated to increase the amount of "annotated" voxels within the bounding boxes. Finally, a weakly supervised regularised loss derived from a Conditional Random Field formulation is used to encourage prediction consistency over homogeneous regions. Extensive experiments are performed on a large open dataset for Vestibular Schwannoma segmentation. textitInExtremIS obtained competitive performance, approaching full supervision and outperforming significantly other weakly supervised techniques based on bounding boxes. Moreover, given a fixed annotation time budget, textitInExtremIS outperforms full supervision. Our code and data are available online.

Key facts. Cited by 21 papers in the map (data through 7 Oct 2026). Builds on 18 of its 26 references that are in the map. First posted on arXiv 1 Jul 2021.

Builds on

18 of 26 references.

References · 18

  1. 2016ScribbleSup: Scribble-Supervised Convolutional Networks for Semantic Segmentationcited by 330
  2. 2020TorchIO: A Python library for efficient loading, preprocessing, augmentation and patch-based sampling of medical images in deep learningcited by 176
  3. 2016Deep Interactive Object Selectioncited by 160
  4. 2017Deep Extreme Cut: From Extreme Points to Object Segmentationcited by 127
  5. 2018On Regularized Losses for Weakly-supervised CNN Segmentationcited by 121
  6. 2017DeepIGeoS: A Deep Interactive Geodesic Framework for Medical Image Segmentationcited by 110
  7. 2017Interactive Medical Image Segmentation using Deep Learning with Image-specific Fine-tuningcited by 110
  8. 2017Extreme clicking for efficient object annotationcited by 80
  9. 2016DeepCut: Object Segmentation from Bounding Box Annotations using Convolutional Neural Networkscited by 79
  10. 2018Constrained-CNN losses for weakly supervised segmentationcited by 62
  11. 2020Bounding boxes for weakly supervised segmentation: Global constraints get close to full supervisioncited by 40
  12. 2020Going to Extremes: Weakly Supervised Medical Image Segmentationcited by 38
6 more
  1. 2020Scribble-based Domain Adaptation via Co-segmentationcited by 34
  2. 2019Automatic Segmentation of Vestibular Schwannoma from T2-Weighted MRI by Deep Spatial Attention with Hardness-Weighted Losscited by 25
  3. 2018Self-Learning to Detect and Segment Cysts in Lung CT Images without Manual Annotationcited by 6
  4. 2020Manual segmentation versus semi-automated segmentation for quantifying vestibular schwannoma volume on MRIcited by 5
  5. 2019Permutohedral Attention Module for Efficient Non-Local Neural Networkscited by 4
  6. 2019Extreme Points Derived Confidence Map as a Cue For Class-Agnostic Segmentation Using Deep Neural Networkcited by 1

Built on by

Cited by year

Citing papers per year: 2021: 1; 2022: 5; 2023: 7; 2024: 4; 2025: 4.

21 papers.

Sort:

  1. 2022Scribble-Supervised Medical Image Segmentation via Dual-Branch Network and Dynamically Mixed Pseudo Labels Supervisioncited by 64
  2. 2022CrossMoDA 2021 challenge: Benchmark of Cross-Modality Domain Adaptation techniques for Vestibular Schwannoma and Cochlea Segmentationcited by 58
  3. 2021Source-Free Domain Adaptation for Image Segmentationcited by 45
  4. 2022PA-Seg: Learning from Point Annotations for 3D Medical Image Segmentation using Contextual Regularization and Cross Knowledge Distillationcited by 16
  5. 2022Scribble2D5: Weakly-Supervised Volumetric Image Segmentation via Scribble Annotationscited by 15
  6. 2023Label-Efficient Deep Learning in Medical Image Analysis: Challenges and Future Directionscited by 14
  7. 2023Weakly-Supervised 3D Medical Image Segmentation using Geometric Prior and Contrastive Similaritycited by 9
  8. 2023SimPLe: Similarity-Aware Propagation Learning for Weakly-Supervised Breast Cancer Segmentation in DCE-MRIcited by 6
  9. 2023Data efficient deep learning for medical image analysis: A surveycited by 6
  10. 2024LNQ 2023 challenge: Benchmark of weakly-supervised techniques for mediastinal lymph node quantificationcited by 6
  11. 2022Annotation by Clicks: A Point-Supervised Contrastive Variance Method for Medical Semantic Segmentationcited by 4
  12. 20243D Vascular Segmentation Supervised by 2D Annotation of Maximum Intensity Projectioncited by 3
9 more
  1. 2024Extreme Point Supervised Instance Segmentationcited by 3
  2. 2023From Sparse to Precise: A Practical Editing Approach for Intracardiac Echocardiography Segmentationcited by 2
  3. 2023Leveraging point annotations in segmentation learning with boundary losscited by 2
  4. 2024OOD-SEG: Exploiting out-of-distribution detection techniques for learning image segmentation from sparse multi-class positive-only annotationscited by 2
  5. 2023Volumetric Medical Image Segmentation via Scribble Annotations and Shape Priorscited by 1
  6. 2025EAUWSeg: Eliminating annotation uncertainty in weakly-supervised medical image segmentationcited by 1
  7. 2025DEPT: Deep Extreme Point Tracing for Ultrasound Image Segmentationcited by 0
  8. 2025Label tree semantic losses for rich multi-class medical image segmentationcited by 0
  9. 2025Uncertainty-Aware Extreme Point Tracing for Weakly Supervised Ultrasound Image Segmentationcited by 0