MartinK84 / NonCrtGRAPPA

A generalized GRAPPA algorithm that works for arbitrary (non-Cartesian) readout trajectories

Non-Cartesian GRAPPA

Reference implementation of:
A GRAPPA algorithm for arbitrary 2D/3D non-Cartesian sampling trajectories with rapid calibration
doi: 10.1002/mrm.27801

cite as:

@article{Luo2019GRAPPA,
author = {Luo, Tianrui and Noll, Douglas C. and Fessler, Jeffrey A. and Nielsen, Jon-Fredrik},
title = {A GRAPPA algorithm for arbitrary 2D/3D non-Cartesian sampling trajectories with rapid calibration},
journal = {Magnetic Resonance in Medicine},
doi = {10.1002/mrm.27801},
url = {https://onlinelibrary.wiley.com/doi/abs/10.1002/mrm.27801},
eprint = {https://onlinelibrary.wiley.com/doi/pdf/10.1002/mrm.27801}
}

General comments

setup_NonCrtGRAPPA.m does configurations.
Demos are provided in ./demo.

# Parameter naming
P:    Image;
kP:   Fourier transform of P;
PS:   Coil images, i.e., P modulated with sensitivity maps;
kPS:  Fourier transform of PS, channel-wise;
ukPS: Under-sampled kPS, zero-filled;
fkPS: Full kPS, either fully sampled or reconstructed.

This repo has included binary test data files for basic accessibility in certain regions.
Future binary data files will be available on https://drive.google.com/drive/folders/1FxF5jcMhL8Z2IzB-i__mb4Q3wDOmi20R.

Dependencies

• MIRT, http://web.eecs.umich.edu/~fessler/irt/fessler.tgz.

Compatible MATLAB versions

For MATLAB < R2018a, if LAPACK library is not correctly linked when building LS_fft_mex.c, one would need an additional flag 'lmwlapack' as:

mex LS_fft_mex.c -lmwlapack