Update: June 10, 2019. The software available from this page has been updated from Version 2.0 to Version 2.1. This new version makes a minor change to substantially reduce memory usage and improve computational efficiency for algorithm 4 (the multiplicative half-quadratic algorithm that uses FFTs with approximation). The original Version 2.0 is still available on request, although we encourage the use of Version 2.1.
This page provides a new MATLAB software implementation for Low-Rank Modeling of Local k-Space Neighborhoods (LORAKS), as described in the technical report:
T. H. Kim, J. P. Haldar.
LORAKS Software Version 2.0: Faster Implementation and Enhanced Capabilities.
University of Southern California, Los Angeles, CA,
Technical Report USC-SIPI-443, May 2018.
[link]
Over the past several years, our research group has been developing a novel structured low-rank matrix modeling framework for magnetic resonance (MR) image reconstruction that we call LORAKS. In the spirit of reproducible research, we had previously released a public open-source software implementation of LORAKS-based image reconstruction in 2014. In the new technical report referenced above (and supplementary material available for download below), we provide access to an updated open-source software release that includes many of the new LORAKS developments that have occurred since 2014, including substantially-faster algorithms and a variety of new formulations of the inverse problem. The download also includes examples of real in vivo MRI k-space data corresponding to human brain images.
An example reconstruction result is shown below, in which we have used P-LORAKS to reconstruct MRI brain data from highly-accelerated calibrationless 4-channel parallel imaging data.
Gold Standard | Calibrationless Sampling |
Zero-filled | P-LORAKS |
The software and data can be downloaded using the form below: