Apply the contrast-agnostic segmentation model

[valosekj]

The CSV files in this repository were generated based on the spinal cord segmentations obtained using sct_deepseg_sc -c t2 (QCed and corrected if necessary) using this script.

Since we now have the contrast-agnostic segmentation model (sct_deepseg -task seg_sc_contrast_agnostic), we can regenerate the segmentations (and CSV) files with a model less sensitive to a specific contrast.