Refractor derivatives naming convention

[sandrinebedard]

Description

Since we have now clarified the naming of the labels in the derivatives, we need to update the derivatives name in spine generic

Suggestions

• sub-amu01_T2star_seg-manual.nii.gz --> sub-amu01_T2star_label-SC_seg.nii.gz
• sub-amu01_T2w_labels-disc-manual.nii.gz --> sub-amu01_T2w_label-discs_dlabel.nii.gz
• sub-amu01_T1w_softseg.nii.gz --> sub-amu01_T1w_label-SC_softseg.nii.gz
• sub-amu01_T1w_softseg_bin.nii.gz --> sub-amu01_T1w_desc-softseg_label-SC_seg.nii.gz (to be discussed)
• sub-amu01_T1w_csfseg-manual.nii.gz --> sub-amu01_T1w_label-CSF_seg.nii.gz
• For discs warped from T2w: sub-amu01_T2w_label-discs_desc-warp_dlabel.nii.gz
• _RPI_r_labels-manual --> space-other_T2w_label-vertebrae_desc-C2C5_dlabel
• gmseg--> label-GM_seg

Note: do we want to add the discs labels from warping from T2w for the other contrasts inside the derivatives? (as we do and need for the contrast-agnostic project)

Related issues

156

154

141

TODO

• Update all preprocessing scripts with new name convention

[sandrinebedard]

tagging @jcohenadad @valosekj @mguaypaq @NathanMolinier

[jcohenadad]

Looks good to me! (i'm copy/pasting in case the issue text changes in the future)

Note: do we want to add the discs labels from warping from T2w for the other contrasts inside the derivatives? (as we do and need for the contrast-agnostic project)

I would say yes, so that the contrast-agnostic model can more easily be reproduced, no?

[naga-karthik]

so that the contrast-agnostic model can more easily be reproduced, no?

Yes! We required the disc labels for vertebral labeling and the C2-C3 CSA computation via sct_process_segmentation. Hence, adding the disc labels would be good!

[sandrinebedard]

@NathanMolinier

We have some disc labels that are warped from the T2w, but we want to add them in the derivatives 2 (BUT THEY MAY NOT BE PERFECT) and we do not want to be confused between them and manual disc labels, do you have a suggetsion?

[jcohenadad]

We should also add the space-other field https://github.com/spine-generic/data-multi-subject/issues/156

[sandrinebedard]

I noticed that the disc labels in the derivatives are in the original space LPI, should I reorient them and add space-other to be consitent? or do we keep them in the original space?

[NathanMolinier]

We should also add the space-other field #156

Or space-warp but we need to add this to the JSON sidecar as well to say that the image was warped from another contrast

[NathanMolinier]

I noticed that the disc labels in the derivatives are in the original space LPI, should I reorient them and add space-other to be consitent? or do we keep them in the original space?

Therefore, do we need to store 2 versions of the discs labels ?

[sandrinebedard]

reorientation is a very simple step, I don't think we need 2 versions (in the contrast-agnsotic pipeline, sct_label_vertebrae reorients in any case)