Script order

• Many of the .py files use helper functions defined in utils.py
• Most of these helper functions just print out useful information, or they help in retrieving common files across the different patient directories (e.g. readANTS and get_common_files)

---

1. resize.m

• This is a Matlab script that resizes the MRI mask to match the size of the CT mask.
• The dim (number of voxels in the xyz directions) and the pixdim (voxel spacing; size of voxel in the xyz directions), are matched to what values they have in the CT mask's header file. This is done as a consequence of the imresize3 function in Matlab.
• After resizing, transfer the local imgs/ directory into the server
• One can check the header file in ITK-SNAP through Tools > Image Information, to confirm that the resize changes were applied.

2. register.sh

• Runs all these scripts:
  1. cleanup.sh
  • This is an optional script that cleans up any files generated from previous registration attempts.
  • Used primarily when testing registration.
  2. reorient.py
  • At this stage, the MRI masks closely match the size of the CT masks, but they just need to be reoriented (and translated, and deformed using the subsequent scripts).
  • This script will modify the affine matrices of the CT masks to have RAS axis codes, and the affine matrices of the MRI masks to have SRA axis codes.
  3. register.py
  • This is the first phase of registration. The resized and reoriented MR masks are aligned to match the CT masks by translation.
  • The type_of_transform parameter was set to Similarity.
  • In the second and final phase of registration, the resized, reoriented and translated MR masks undergoes a deformation so that the boundary of the MR mask closely matches the boundary in the CT mask.
  • The type_of_transform parameter in the second stage was set to SynAggro.
  4. unpack.py
  • This script loads the warped MRI images from the serialized output of the similarity registration in the previous step.
  • The warped MRI images are saved in the patient directories (e.g. PIm0028) that they originally came from.
  5. unzip.sh
  • This script just unzips the warped MRI images generated in the registration process.
  • The file extension of the warped MRI images changes from .nii.gz to .nii
  6. warp_vent.py
  • Applies the forward transformations from the registration (of the MRI mask onto the CT mask) to the ventilation image as well

First registration phase finished on 11/5/2024

git stuff

1. git init makes this a master branch by default
2. gite remote add origin <repo>
3. git checkout main
4. git status to ensure that the current branch is main
5. git pull origin main --allow-unrelated-histories to pull stuff in main that is not in master because the default branch is master
6. git push origin main this puts all the stuff in master into main
7. git push origin --delete master deletes master; now we just have main`

Current problems

• ???

Current tasks

• Figure out how VDP is calculated in duke pipeline

Masking a ventilation image with lobar mask (example)

mask.shape
vent_mask_volm.shape
masked = apply_mask(mask=mask, data=vent_mask_volm)
nib.save(img=nib.nifti1.Nifti1Image(masked, vent_mask_imag.affine), filename="___.nii")