Application
The ISMRM Raw Data format is proposed as a common MR raw data format to promote collaborations and reproducible research by facilitating the sharing of data and data processing tools.
Contributors
Michael S. Hansen1, Souheil J. Inati2, Joseph D. Naegele2, Nicholas R. Zwart3, Kaveh Vahedipour4
The ISMRM Raw Data (ISMRMRD) format is proposed as a common MR raw data format to facilitate the sharing of data and data processing tools. It is composed of the data fields required to reconstruct images from magnetic resonance experiments. C/C++, Python and MatLab application interfaces have been developed to translate MR data from the different MR vendors into this open data format.
The data is stored in HDF5 format combining a mix of flexible and fixed data structures in the next way:
- A flexible document that can contain an arbitrary number of fields in an XML format. This document stores parameters that are specific to the experiment.
- A raw data section that holds the data acquired from the experiment.
- Additional post-processed data, such as reconstructed images and coil sensitivity maps, can be stored in the HDF5 header.
Other than the raw data necessary to reconstruct the images, the ISMRMRD format does not explicitly define how extra data should be stored, but leaves it open for the user to add variables and data as dictated by each application.
This standard has been developed by the contribution of plural MR community members, and remains in constant development as the vendors release new software versions. The members of the research community are invited to contribute with future ISMRMRD versions and associated conversion tools, promoting collaborations and reproducible research.
Publications
Affiliations
1National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, Maryland, USA.
2National Institute of Mental Health, National Institutes of Health, Bethesda, Maryland, USA.
3Keller Center for Imaging Innovation, Barrow Neurological Institute, Phoenix, Arizona, USA.
4Institute of Neuroscience and Medicine-4, Juelich, Germany