Create a processed version of the dataset

[sandrinebedard]

Context

We want to use spine-generic/data-multi-subject dataset for training of sct_deepseg. However, we have encountered some issues: the SC segmentations are generated with the resampled and reoriented data.

Also, the SC segmentation is not done for all available contrasts in the dataset, we want to add those.

Suggestion

A solution for this issue is to create a processed version of the spine-generic/data-multi-subject like we did for UK-Biobank.

• Create spine-generic-processed with preprocessed data
• Remove suffix _RPI_r on the data and in the derivatives --> e.g. of what we have now sub-brno03
• Add the automatic segmentations that are manually QC-ed to the derivatives