DICOM phase image scaling

[Lestropie]

For Siemens data, phase images appear in the range [-4096, 4094]. Therefore, before such an image can be combined with the corresponding magnitude image to form a complex image via mrcalc -polar, it must be explicitly rescaled to radians units.

I however recall seeing another piece of software being used locally in the lab, processing multi-echo GRE data for SWI, where this explicit scaling was not necessarily. The software was seemingly identifying somewhere else within the DICOM headers relevant intensity scaling information defined by the vendor. It would be much more elegant if we could do the same for such data.