Our Magnetic Resonance Imaging software resides in the GitHub repository https://github.com/MagneticResonanceImaging/MRIReco.jl. This repository contains Julia packages for the whole MRI reconstruction pipeline including dataloading, data-preprocessing, and the actual image reconstruction task. Base functionality underlying all of these packages are provided by the package MRIBase.jl. This package contains implementations for commonly used objects that are central to MRI (e.g. objects describing the measured MRI data or the sampling trajectory used during the measurement). Based on this functionality, MRIFiles.jl provides support for the reading and writing of raw MRI data. Importantly, we provide read- and write support for the open-source ISMRMRD format and read support for raw MRI data from the vendor Bruker.

An important building block of all MRI reconstruction algorithms are the imaging operators, which describe the relationship between image and the measured MRI signal. To address this need, the package MRIOperators.jl provides efficient implementations of commonly used MRI operators. Notably, we support not only Cartesian and non-Cartesian Fourier measurement operators, but also operators taking into account inhomogeneities of the main magnetic field. Another important aspect of MRI reconstruction is the estimation and compression of the receive-coil sensitivity maps. This kind of functionality is contained in the package MRICoilSensitivities.jl.

Finally, MRIReco.jl contains implementations of commonly used MRI image reconstruction algorithms. An emphasis of this package is to provide an interface that can be easily extended and customized. This makes it easy, to incorporate custom code and thus facilitates the development of new algorithms.

Contact

• Tobias Knopp