Many members of the ISMRM community develop customized software tools to solve problems in various aspects of MR sequence design, image reconstruction and data processing. The MR-Hub offers a platform where researchers can share their software solutions with the rest of the community - hopefully making more people aware of existing tools, allowing others to solve their own problems more rapidbly by building on existing solutions. We encourage all members of the ISMRM community to follow the spirit of reproducible research, and consider making the code behind their publications available to share.

This is just the start of an effort to upgrade the old MR Unbound site. We are just showing a few examples to start the process, but we encourage anyone to submit new or existing software to be shared and indexed here.

All software here is tested independently to help ensure a smooth experience and to provide feedback for developers.

Detailed submission instructions:

You are invited submit descriptions of open source software packages you have written. The software packages will be evaluated by at least one of your peers and posted on the ISMRM MR-Hub site along with the impressions from the evaluation. During the evaluation process, you will have the chance to work with the evaluator(s) to describe the software and ensure it worked as intended during evaluation.

To ensure that your software package gets the best possible exposure, it is recommended to follow the following guidelines:

  1. We strongly recommended that you upload your source code to a well-known platform, e.g. GitHub or Bitbucket, as these platforms encourage community contributions and discussion. For this reason we recommend against hosting, for example, a zip file of the whole package on your institution website.
  2. If appropriate, provide a working package that can be tested without having to build the software. Scripting packages (Matlab, Python, etc.) usually run out of the box, but please make sure that dependencies are included. For compiled code or software environments that may be complicated to configure, please consider providing a virtual machine image, Docker container, Anaconda environment, or a similar solution that would allow potential users to get a version of the software up and running as quickly as possible to try it out. Make sure that any external dependencies are used only when absolutely necessary (e.g., for licensing reasons) and are well described.
  3. Provide meaningful examples that can run on average personal computers along with instructions that include descriptions of input and expected outputs. A tutorial walk-through (including installation) is a good way to demonstrate the software.

What to include in your submisson?

For the summary featured on the MR-Hub page:

To make the review process as smooth as possible, please make sure that by following the links you have submitted for the MR-Hub summary it will be easy to find:

Please then send all this information to MRHub@ismrm.org!  


Software packages:

Reconstruction toolbox and programming library for parallel imaging and compressed sensing available for Linux, Mac OS X, and Windows.
Principal developers: Martin Uecker, Jon Tamir and Frank Ong

The Berkeley Advanced Reconstruction Toolbox (BART) is a free and open-source image- reconstruction framework for Computational Magnetic Resonance Imaging. It consists of a programming library (for C/C++) and a toolbox of command-line programs. The library provides common operations on multi-dimensional arrays, (non-uniform) Fourier and (divergence-free) wavelet transforms, as well as generic implementations of iterative optimization algorithms (CG, IST, FISTA, ADMM, IRGNM, ...) supporting various types of regularization terms (l2, l1-wavelet, total-variation, low-rank, multi-scale low-rank ...) using pre-defined functions. Parallel computation on multi-core systems and on Graphical Processing Units (GPUs) is supported. The command-line tools provide direct access to basic operations on multi-dimensional arrays as well as efficient implementations of many calibration and reconstruction algorithms for parallel imaging and compressed sensing (SENSE, ESPIRiT, NLINV, SAKE, ...).

Journal Publications using BART:

Martin Uecker, Peng Lai, Mark J. Murphy, Patrick Virtue, Michael Elad, John M. Pauly, Shreyas S. Vasanawala, and Michael Lustig. ESPIRiT - An Eigenvalue Approach to Autocalibrating Parallel MRI: Where SENSE meets GRAPPA. Magn Reson Med 2014; 71:990-1001.

Hollingsworth KG, Higgins DM, McCallum M, Ward L, Coombs A, Straub V. Investigating the quantitative fidelity of prospectively undersampled chemical shift imaging in muscular dystrophy with compressed sensing and parallel imaging reconstruction. Magn Reson Med 2014; 72:1610-1619.

Zhang T, Cheng JY, Potnick AG, Barth RA, Alley MT, Uecker M, Lustig M, Pauly JM, Vasanawala SS. Fast Pediatric 3D Free Breathing Abdominal Dynamic Contrast Enhanced MRI with a High Spatiotemporal Resolution, J Magn Reson Imaging 2015; 41:460-473.

Addy NO, Ingle RR, Wu HH, Hu BS, Nishimura DG. High-resolution variable-density 3D cones coronary MRA. Magn Reson Med 2015; 74: 614-621.

Cheng JY, Zhang T, Ruangwattanapaisarn N, Alley MT, Uecker M, Pauly JM, Lustig M, Vasanawala SS. Free- Breathing Pediatric MRI with Nonrigid Motion Correction and Acceleration, J Magn Reson Imaging 2015; 42:407-420.

Athalye V, Lustig M, Uecker M. Parallel Magnetic Resonance Imaging as Approximation in a Reproducing Kernel Hilbert Space, Inverse Problems 2015; 31:045008.

Mann LW, Higgins DM, Peters CN, Cassidy S, Hodson KK, Coombs A, Taylor R, Hollingsworth KG. Accelerating MR Imaging Liver Steatosis Measurement Using Combined Compressed Sensing and Parallel Imaging: A Quantitative Evaluation, Radiology 2016; 278:245-256.

Cheng JY, Hanneman K, Zhang T, Alley MT, Lai P, Tamir JI, Uecker M, Pauly JM, Lustig M, Vasanawala SS. Comprehensive Motion-Compensated Highly-Accelerated 4D Flow MRI with Ferumoxytol Enhancement for Pediatric Congenital Heart Disease, J Magn Reson Imaging, Epub (2015)

Tamir JI, Uecker M, Chen W, Lai P, Aleey MT, Vasanawala SS, Lustig M. T2-Shuffling: Sharp, Multi-Contrast, Volumetric Fast Spin-Echo Imaging, Magn Recon Med, Epub (2016).

Martin Uecker and Michael Lustig, Estimating Absolute-Phase Maps Using ESPIRiT and Virtual Conjugate Coils, Magnetic Resonance in Medicine, Epub (2016)

General-purpose, open source, medical imaging reconstruction framework.
Principal developers: Michael Hansen, Thomas Sørensen

The Gadgetron is an Open Source, general-purpose medical imaging reconstruction framework written primarily in C++. It consists of two main components: 1) a set of versatile toolboxes for image signal processing, and 2) a modular, high performance framework for streaming data processing. The streaming framework uses a client server model where the reconstruction job is performed on a server and the client is responsible for sending data and receiving imaging. Clients can be stand-alone command line clients or an imaging device could serve as a client. The framework supports easy prototyping with scripting languages such as Python while supporting transparent integration in clinical work flows. For MRI applications, the Gadgetron supports the vendor independent ISMRM Raw Data format and it comes with high performance image reconstruction pipelines for many standard MRI sequences. Example applications include Cartesian and non-Cartesian parallel imaging, non-linear reconstruction and motion correction. Several performance critical components such as the non-Cartesian Fourier transform have been implemented on GPUs for improved performance. The framework can be used as a complete reconstruction package or the toolboxes can be leveraged in other applications.

Hansen MS, Sørensen TS. Gadgetron: An Open Source Framework for Medical Image Reconstruction.
Magn Reson Med. 2013 Jun;69(6):1768-76.

C++ CUDA accelerated non-uniform FFT for arbitrary 2D/3D data with direct Matlab interface on Windows and Linux
Principal developers: Andreas Schwarzl, Florian Knoll

The computational expensive non-uniform FFT is implemented in C++ by utilizing the NVIDIA CUDA architecture, in order to speed up the execution of the 2D/3D multi-coil Gridding step for arbitrarily shaped k-space data.

The software offers a ready-to-use interface to Matlab, similar to the well-known NUFFT Toolbox by Fessler et al., in order to seamlessly enhance the preferred Matlab prototyping process at an early stage.

The gpuNUFFT supports the growing importance of non-Cartesian 2D/3D trajectories in the context of iterative image reconstruction, parallel imaging, and compressed sensing by reducing the long computation times during NUFFT operations. Depending on the data size and shape, speedups up to a factor of 40 are possible.

F. Knoll, A. Schwarzl, C. Diwoky, and D. Sodickson, “gpuNUFFT - an open-source GPU Library for 3D Gridding with direct Matlab interface”, in Proceedings of the 22th Annual Meeting of ISMRM, Mailand, 2014 , p4297 Available: http://submissions.miracd.com/ismrm2014/proceedings/

Vendor-neutral MRI raw data format based on standard developed by a subcommittee of the ISMRM Sedona 2013 workshop
Principal developers: See full list of developers

A prerequisite for sharing MRI reconstruction software is a data format that describes typical MRI raw datasets in a vendor neutral manner. The ISMRM Raw Data format is a proposal for such a common MR raw data format, which promotes algorithm and data sharing. The file format consists of a flexible header and tagged frames of k-space data. These data elements are stored in an HDF5 file container. Application Programming Interfaces are available for C/C++, MATLAB, and Python. Converters for Bruker, General Electric, Philips, and Siemens proprietary file formats also available (implemented in C++). The proposed raw data format solves a practical problem for the magnetic resonance imaging community. It may serve as a foundation for reproducible research and collaborations. The ISMRM Raw Data format is a completely open and community-driven format, and the scientific community is invited (including commercial vendors) to participate either as users or developers.

Inati SJ, Naegele JD, Zwart NR, Roopchansingh V, Lizak MJ, Hansen DC, et al. ISMRM Raw data format: A proposed standard for MRI raw datasets. Magn Reson Med. 2016; doi:10.1002/mrm.26089

This package is a MATLAB based tool for simulating low-field MRI acquisitions based on high-field acquisition, enables prediction of the minimum field strength requirements for a broad range of MRI techniques.
Principal developers: Weiyi Chen, Ziyue Wu, Krishna S. Nayak

This package is to develop and evaluate a framework for simulating low-field proton-density weighted MRI acquisitions based on high-field acquisitions, which could be used to predict the minimum B0 field strength requirements for MRI techniques.

Given MRI raw data, lower field MRI acquisitions can be simulated based on the signal and noise scaling with field strength. Certain assumptions are imposed for the simulation. This package also contains two examples of proton-density weighted MRI applications that demonstrated estimation of minimum field strength requirements: real-time upper airway imaging and liver proton-density fat fraction measurement. More detailed description of the examples can be found in the published reference.

This package enables prediction of the minimum field strength requirements for a broad range of MRI techniques, and would be particularly useful in the evaluation of de-noising and constrained reconstruction techniques.

Z Wu, W Chen, KS Nayak. "Minimum Field Strength Simulator for Proton Density Weighted MRI." PLoS ONE 11(5): e0154711. May 2016.
http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0154711

Z Wu, W Chen, KS Nayak. "Low-Field Simulation and Minimum Field Strength Requirements." ISMRM Workshop on Data Sampling and Image Reconstruction, Sedona, Arizona, Jan 2016.
http://cds.ismrm.org/protected/Data16/Program/Abstracts/Wu.pdf