FFDR scan issue

[ihnorton]

Given a DTI dataset scanned with PatientPosition as FFDR (Face First Decubitus Right) rather than the common HFS (Head First Supine):

Need to ensure that DWIConvert, as well as tensor estimation and tractography gradient direction usages, consistently account for this non-standard orientation.

[ihnorton]

I also tested this data with other software, and it seems that several others also have incomplete support for "non-standard" scanning position:

• DiffusionToolkit/TrackVis and MITK both silently give strange results
• dcm2nii prints an error message: Siemens/Philips DTI directions require head first supine acquisition (proceeding anyway, the results from TrackVis are strange).