Generalizability of the CNN

[pkashya]

I've been trying to run your fetal brain segmentation algorithm on our recently collected fetal fMRI data using the code on Github.
It seems to run fine on google collab, but the predicted outputs for each volume seem to be incomplete (attached are a few examples on the recent email to s**.r**@donders.ru.nl )...I'm wondering if its because the CNN needs to be re-trained on our data?

[Achillesy]

@pkashya
If you have fmri data, try my segmentation program: https://github.com/Achillesy/Fetal_Functional_MRI_Segmentation, no configuration required.