DICOM information

DICOM is a large and sometimes confusing imaging data format.

In the other pages in this series we try and document our understanding of various aspects of DICOM relevant to converting to formats such as NIfTI .

There are a large number of DICOM image conversion programs already, partly because it is a complicated format with features that vary from manufacturer to manufacturer.

We use the excellent PyDICOM as our back-end for reading DICOM.

Here is a selected list of other tools and relevant resources:

  • Grassroots DICOM : GDCM . It is C++ code wrapped with swig and so callable from Python. ITK apparently uses it for DICOM conversion. BSD license.

  • dcm2nii - a BSD licensed converter by Chris Rorden. As usual, Chris has done an excellent job of documentation, and it is well battle-tested. There’s a nice set of example data to test against and a list of other DICOM software. The MRIcron install page points to the source code. Chris has also put effort into extracting diffusion parameters from the DICOM images.

  • SPM8 - SPM has a stable and robust general DICOM conversion tool implemented in the spm_dicom_convert.m and spm_dicom_headers.m scripts. The conversions don’t try to get the diffusion parameters. The code is particularly useful because it has been well-tested and is written in Matlab - and so is relatively easy to read. GPL license. We’ve described some of the algorithms that SPM uses for DICOM conversion in SPM DICOM conversion.

  • DICOM2Nrrd: a command line converter to convert DICOM images to Nrrd format. You can call the command from within the Slicer GUI. It does have algorithms for getting diffusion information from the DICOM headers, and has been tested with Philips, GE and Siemens data. It’s not clear whether it yet supports the Siemens mosaic format. BSD style license.

  • The famous Philips cookbook: https://www.archive.org/details/DicomCookbook

  • http://dicom.online.fr/fr/dicomlinks.htm