-
Notifications
You must be signed in to change notification settings - Fork 228
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Philips DWI conversion to NRRD and NIFTI #612
Comments
NOTE: The "measurement frame" is the RAS version of the inverse of the dicom ImageOrientationPatient field
|
@tashrifbillah is the developer and maintainer of the NRRD support. He describes the transforms here. He describes tests here and includes Python converters. I tested the NIfTI code with FSL. I include compact Philips validation datasets for several generations of Philips software. The NITRC page contains many more large validation datasets - I also encourage users to read the dedicated document that is linked from that page that describes how to test the conversion of diffusion vectors. |
@neurolabusc Thank you for the information. The links are very useful. I'll reach out to @tashrifbillah. Hans |
@neurolabusc @tashrifbillah Thank you for the python code! I'll review carefully and make queries if I run into any issues. Thank you! Hans |
@hjmjohnson it would be great to have an independent review of the dcm2niix NRRD conversion. |
Describe the bug
Gradient directions have different interpretations for NIFTI and NRRD converted files, but have the same numerical values.
To reproduce
For data collected for a single shell DWI sequence on a :
run
dcm2niix -e n ...
run
dcm2niix -e y ...
and compare the gradients that are written to files.
Expected behavior
Gradient direction values are encoded differently in nhdr and bvec files, so they should have different numerical values. The nifti variant has an implied identity measurement frame. The nhdr file can have a non-identity measurement frame. If the measurement frame in the nhdr file is not identity, then the values in the bvec file should be different from those stored in the nhdr file.
The first gradient (second column) of values stored for fsl format bvec file (Implied identity measurement frame).
First gradient stored in NRRD format header:
Version
dcm2niix -v
Chris Rorden's dcm2niiX version v1.0.20211006 (JP2:OpenJPEG) GCC9.3.0 x86-64 (64-bit Linux)
v1.0.20211006
(same results from today's development build)
Troubleshooting
I reviewed content on https://www.na-mic.org/wiki/NAMIC_Wiki:DTI:Nrrd_format.
https://www.na-mic.org/wiki/File:Coords-RAS.png
** My hypothesis is that the gradients written to the nhdr file have already been adjusted for the measurement frame, and should have an identity measurement frame written to the nhdr file.
The text was updated successfully, but these errors were encountered: