File Organization

Directories should be organized by dataset ID with the prefix dataset, i.e., a directory called dataset_10067.
Tomogram images should be organized by run ID and placed inside dataset directories with the prefix run, i.e., dataset_10067/run_1012.mrc or dataset_10067/run_1012.mha.
Segmentations should be organized by run ID and placed inside dataset directories with the prefix seg, i.e., dataset_10067/seg_1012.mha.

Pipeline

Download seg_setup.sh, install ITK-SNAP.
Run the script seg_setup.sh by typing source seg_setup.sh where the script is located.
Be sure that ITK-SNAP is installed and callable from the terminal with the command itksnap. In Linux this seems to require editing $PATH to include the itksnap executable included in the downloaded directory.
Be sure the current python environment has the necessary packages installed (like mrcfile). This is easy with a Python virtual environment. On the Mac, where setup is complete, call source py_research/bin/activate.
Navigate to folder containing .mrc data
Call segment [path/to/myfile.mrc] to convert myfile.mrc to a .mha file and open in in ITK-SNAP
Segment the image as desired
When you are done segmenting, in the program, do the following: - Save the segmentation by selecting "Segmentation -> Save Segmentation Image..." and use the .mha (MetaImage) filetype, with whatever filename you choose. - Save the labels by selecting "Segmentation -> Label Editor -> Actions... -> Export", with whichever filetype or filename you choose.
Close itksnap
In the terminal, the segmenatation and labeling you have just created are now saved in the SegData folder.
In the folder in which the original .mrc file came from, call to_julia SegData/[mysegmentation.mha] to convert the .mha segmentation data to a Julia array, which is saved in a .jld2 (JLD2) file.