This PR adds a script compute_csa.sh to compute CSA on the prediction mask of the trained models.
The script can be used with sct_run_batch as described in the README. The input of the script is the clean processed dataset (data_processed_clean) resulting from the preprocessing. The path of the prediction masks is included as an extra script argument.
Breifly, if the subject has a prediction mask for the corresponding contrast, the discs labels (from data_processed_clean/derivatives) are used to created a labeled segmentation. CSA is computed at C2-C3 verterbral levels using the prediction mask. One .csv file is created per contrast.
Description
This PR adds a script
compute_csa.sh
to compute CSA on the prediction mask of the trained models.The script can be used with
sct_run_batch
as described in the README. The input of the script is the clean processed dataset (data_processed_clean) resulting from the preprocessing. The path of the prediction masks is included as an extra script argument.Breifly, if the subject has a prediction mask for the corresponding contrast, the discs labels (from data_processed_clean/derivatives) are used to created a labeled segmentation. CSA is computed at C2-C3 verterbral levels using the prediction mask. One .csv file is created per contrast.