You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Currently, the manual rootlets GT segmentations (n=36; see D5.tsv) used to train the latest model M5 are stored locally on duke (~/duke/projects/ml_spinal_rootlets/datasets/Dataset012_M5) organized according to the nnUNet format:
Note that for five testing images (imagesTs) we have available reference STAPLE segmentations (labelsTs) and also manual segmentations from 4 raters. These segs are already BIDSified (/duke/projects/ml_spinal_rootlets/datasets/inter-rater_variability):
we also need to find a way to distinguish rater1, rater2, staple, etc. Maybe @NathanMolinier has some ideas
To differentiate between raters and staple, I think that we could use the entity desc-XXX (desc-rater1, desc-staple etc.). See BIDS specification.
Also, if we want to clarify the meaning of these desc entities, we could also add a file descriptions.tsv at the root of the derivative folder as shown on our intranet here.
Currently, the manual rootlets GT segmentations (n=36; see D5.tsv) used to train the latest model M5 are stored locally on duke (
~/duke/projects/ml_spinal_rootlets/datasets/Dataset012_M5
) organized according to the nnUNet format:TODO
labelsTr
) from nnUNet format to BIDS using convert_nnUNetV2_to_bids.pysub-001_acq-sag_T2w_label-rootlets_dseg.nii.gz
Note that for five testing images (
imagesTs
) we have available reference STAPLE segmentations (labelsTs
) and also manual segmentations from 4 raters. These segs are already BIDSified (/duke/projects/ml_spinal_rootlets/datasets/inter-rater_variability
):details
The text was updated successfully, but these errors were encountered: