MoCoLoR

Motion-compensated low-rank reconstruction for simultaneous structural and functional UTE lung MRI

Reference

Tan F, Zhu X, Chan M, Zapala MA, Vasanawala SS, Ong F, Lustig M, Larson PEZ. Motion-compensated low-rank reconstruction for simultaneous structural and functional UTE lung MRI. Magn Reson Med. 2023. doi: 10.1002/mrm.29703

Dependency

Tested with Python 3.10.6:

• numpy (>=1.21)
• sigpy (tested ==0.1.16)
• antspyx
• h5py
• pydicom
• cupy
• numba
• tqdm
• scipy

Example Usage

# convert ute data when using the uwute sequence
# uses the output 'MRI_Raw.h5' of pcvipr_recon_binary created using -export_kdata flag
python convert_uwute_npy.py ${file_dir} ${file_dir}

# run xd reconstruction
python recon_xdgrasp_npy.py ${file_dir} --res_scale 1 --vent_flag 1

# run lor reconstruction
python recon_lrmoco_npy.py ${file_dir} --lambda_lr 0.05 --vent_flag 1 --mr_cflag 0

# run mocolor reconstruction
python recon_lrmoco_vent_npy.py ${file_dir} --lambda_lr 0.05 --vent_flag 1 --reg_flag 1

# export DICOMs
python dicom_from_npy.py ${file_dir} ${orig_dicom_dir}

# rm temp files
rm bcoord.npy bdcf.npy bksp.npy _M_mr.npy