Published July 7, 2021 | Version v2
Other Open

Data and Pre-Trained Models from "Inter Extreme Points Geodesics for End-to-End Weakly Supervised Image Segmentation"

Authors/Creators

  • 1. King's College London

Contributors

Researcher:

  • 1. King's College London

Description

Data and pre-trained models from InExtremIS: Dorent, et al: ""Inter Extreme Points Geodesics for End-to-End Weakly Supervised Image Segmentation". MICCAI (2021) 

Description of the data:

  • full_annotations.zip: Manual segmentations performed in consensus by the treating neurosurgeon and physicist.
  • extreme_points.zip: Manual extreme points performed by a biomedical engineer and simulated extreme points generated using the manual segmentations
  • precomputed_geodesics.zip: Pre-computed geodesics using either the image gradient information only or the image gradient information and the Euclidean distance.
  • models.zip:  contains the pre-trained models used in the experiments.

Files

extreme_points.zip

Files (260.4 MB)

Name Size Download all
md5:1896933664210d5484ba897df5b66c4e
29.0 MB Preview Download
md5:4341befe8740df47fc3c824d850d3c09
357.7 kB Preview Download
md5:55a30bd42011932fa6f692de86c37e99
229.6 MB Preview Download
md5:2910854c6a0a5b920954fc5b0bf0f4ab
1.5 MB Preview Download

Additional details

Related works

Is compiled by
Conference paper: arXiv:2107.00583 (arXiv)