I'm running release 1.9.10 of BXH_XCEDE tools on Solaris 10 built from source code, gcc 3.4.6.

I seems that dicom2xcede does not detect the frequency direction when the input is Philips enhanced MR format. The InplanePhaseEncodeDirection tag is in an MR FOV/Geometry sequence for this format and seems to be found by bxh_dicom.cpp (it does appear as an acqParam). But it is not used to set the frequency direction acquisition parameter. As a result, fmriqa_phantom skips the ghostiness analysis step.

Note that there's an oddity in the DICOM standard (PS 3.3-2009). It shows for tag "Inplane Phase Encoding Direction" (0018,1312) when in the MR FOV/Geometry sequence the enumerated values are ROW, COLUMN or OTHER. When it's an MR Image Module Attribute, the values for the same tag are ROW or COL. Philips sets it to "COLUMN" when that is the direction so the strcmp used in bxh_dicom.cpp will need to be modified accordingly.

A stability phantom acquisition in Philips enhanced MR DICOM format (from Philips Achieva 3T R3.2.1) can be downloaded from http://godzilla.kennedykrieger.org/~jgil... It is 80 x 80 x 20 slices x 300 time points.

Thanks for any help you can provide.

Joe

Yes, that fixes it. I have just tested it out on your sample data and
some other Philips data. I am in the process of pushing out a new
release now. The new release will have the fix. Thanks,

-syam