People are starting to use the multi subject spine-generic dataset and train segmentation models with it. The problem is: each group is manually correcting the produced segmentations/labels. Since we already did that, we should publish the processed dataset, so it can be used (and cited) by others.
What derivatives to include:
the ones used for the contrast-agnostic segmentation (hard SC segmentation for each contrast)
GM segmentation
Vertebral labeling (and disc labeling?)
Todo (please modify):
[ ] Upload to Amazon S3 server
[ ] Create Zenodo entry to get DOI
[ ] Write short paper
[ ] Submit to journal (suggestions of journals: XXX)
People are starting to use the multi subject spine-generic dataset and train segmentation models with it. The problem is: each group is manually correcting the produced segmentations/labels. Since we already did that, we should publish the processed dataset, so it can be used (and cited) by others.
What derivatives to include:
Todo (please modify):