abrys
commented
1 year ago Dear Ilona,
We do have quite some experience with data from adni however never encountered any such issue. What you are describing sounds like the data was at some point wrongly converted to multiframe and back or something like that.
To my knowledge GE does not export any multiframe dicoms and the "PerFrameFunctionalGroupsSequence" should only be present in multiframe dicoms.
As we use dicom2nifti in our clinical applications / medical device software we probably will not fix such conversion issues as typically we would want to avoid running on data that has been tampered with.
Could you perhaps point me to which specific case you are encountering the issues with? Dataset/Patient/Study I will see if I can implement a fix without breaking any other conversions.
Kind regards, Arne
Hi,
I tried converting an MP-RAGE from a GE scanner (part of the ADNI dataset), and using your package as it is, I get "AttributeError: module 'dicom2nifti' has no attribute 'settings'", the conversion failing.
I managed to trouble-shoot the error and what I found was that the culprit is the function "is_multiframe_dicom(dicom_input)" in common.py. It wrongly concludes that my data are 4D, as the header file (of my 3D anatomical image) possesses the field "PerFrameFunctionalGroupsSequence", even though it is empty.
When I comment out the following two lines from common.py, the conversion works:
if "PerFrameFunctionalGroupsSequence" in header: return True
Not sure that this is the best solution to deal with it, I now do get warnings "Removing duplicate slice from series".
I hope you find a way to prevent this issue in the future, thanks very much for the package,
Ilona