The x value you report is the index of the voxel. My understanding is that if
image data is stored so that the determinant of the voxel-to-world mapping is
negative, then MRIcron will flip it for its internal use so that it is able
to work with a voxel-to-world mapping with a positive Jacobian determinant.
This also results in the indexing being reversed if you save an ROI image
with MRIcron. The voxel-to-world transform of the indexing-reversed file
will encode the correct orientation - so that all NIfTI compliant software
should be able to deal with the data correctly. However, there may be some
occasional problems if software assumes that images are aligned
voxel-for-voxel (I had my own experience of this a few weeks ago with a bit
of ad hoc code).
Is this right Chris?
Best regards,
-John
On Friday 19 September 2008 10:57, cyril pernet wrote:
> Hi there,
>
> well as as the subject states I have some left/right troubles ...
>
> I made some images in matlab using spm functions and always load the
> defaults before reading/writing. Now if I read my image in SPM I got my
> voxel of interest in mm at 26 -64 -28 and the same with MRICron and the
> image displays in both SPM and MRICron are on the right side .. but
> the voxel coordinate is x=33 in SPM and x=59 in MRICron ..
>
> dim = [-1 0 0; 0 1 0; 0 0 1]
> defaults.analyze.flip = 1
>
> Does anyone have an idea of what is happening? I beleive SPM voxel
> coordinates are the right ones ...
>
> Cheers
> Cyril
|