SCT Concepts

This section documents some SCT concepts and other useful things to know when using it.

Template / Atlas

Tips & Tricks

Segmentation of the Spinal Cord

SCT provides several tools to perform SC segmentation:

The latter one, using a deep learning model, is giving the best results on most cases, but is not configurable.

The former one is the fallback tool. It has lots of options that can be useful when segmenting tricky volumes. You may use it if sct_deepseg_sc is performing worse results than sct_propseg with default parameters.

Segmentation of GM/WM

SCT provides several tools to perform GM/WM segmentation:

The latter one, using a deep learning model, is giving the best results on most cases.

The former one is the fallback tool.

Temporary Directories

Many SCT commands will create in temporary directories to operate, and there is an option to avoid removing temporary directories, to be used for troubleshooting purposes.

If you don’t know where your temporary directory is located, you can look at: https://docs.python.org/3/library/tempfile.html#tempfile.gettempdir

Matlab Integration on Mac

Matlab took the liberty of setting DYLD_LIBRARY_PATH and in order for SCT to run, you have to run:

setenv('DYLD_LIBRARY_PATH', '');
Prior to running SCT commands. See
https://github.com/neuropoly/spinalcordtoolbox/issues/405

Quality Control

Some SCT tools can generate Quality Control (QC) reports. These reports consist in “appendable” HTML files, containing a table of entries and allowing to show, for each entry, animated images (background with overlay on and off).

To generate a QC report, add the -qc command-line argument, with the location (folder, to be created by the SCT tool), where the QC files should be generated.