Some questions about the data

Some questions about the data  

  By: mwilms on July 19, 2023, 9:30 p.m.

Hi,

Thanks for organizing this challenge! We have a few of questions about the training data:

  • There appear to be some cases (~7) where the label image is blank/does not contain any segmented voxels. That's correct, right? Those are the cases mentioned as "(0% of the brain injured) " in Tab. 1 of the paper 1, right?
  • Will the clinical data and the atlases mentioned in the paper be made available for the challenge?

Thanks, Matthias

Re: Some questions about the data  

  By: baorinawork on July 20, 2023, 2:45 a.m.

Hi Matthias,

(1) Yes, those cases are mild HIE cases, which don't show explicit lesions in neonatal MRIs (labels are all 0s).  
      Those cases are mentioned as "(0% of the brain injured) " in Tab. 1 of our paper.

(2) For this year's challenge, we expect participating teams to use the ADC maps, Z_ADC maps or combine them to predict HIE lesions. 
 Therefore, we didn't release the clinical data and atlases for now.  After this year's challenge or next year's challenge, we will release that information.

Thanks, Rina

 Last edited by: baorinawork on Aug. 15, 2023, 12:58 p.m., edited 1 time in total.