Hi Kate,
> I used SPM5b, revision 218, DICOM import function to convert
> 2 Siemens DICOM volumes to nifti. One volume was an MPRAGE, 1 slice
> per dcm file, the other was a functional image in DICOM mosaic format.
>
> Why did the conversion set the sform ? I thought there would only
> be a qform setting indicating the "scanner-anatomical" coordinates.
I was never all that keen on having two sets of transformation matrices in the
header, so I chose to make them both identical as far as possible. Wherever
possible, SPM5 takes the orientation information from the sform. If there is
no sform, then the qform is used. There are some affine transforms encoded
by DICOM headers that can not be represented by the qform (e.g. some CT
images where the gantry is tilted).
I figured that by replicating the same matrices in both the sform and qform,
then there would be less chance of incompatibilities arising.
> Is there an additional "put it in SPM default space" translation, or did
> it convert to a Talairach space, and in any case wouldn't the sform code
> be something besides 1 ?
For now, I've chosen to make the sform and qform codes 2 after any kind of
reorienting. After any such procedure, the qform and sform codes will remain
identical (as far as possible).
All the best,
-John
|