On Tue, 15 Nov 2005, Jim Lee wrote:

> When I acquire images of a subject whose head is slightly tilted, I
> tilt the image stack to compensate, so the subject's head is lined up
> with the image stack (Scanning on a Siemens 3T Trio).
> 
> If I convert the dicom images using SPM2's Dicom toolbox, the tilt of
> the image stack is faithfully preserved. I would rather have a straight
> image stack, which would compensate for the subject's tilted head.
> 
> Has anyone modified spm_dicom_convert to do this? I tried a simple hack
> (replacing hdr{i}.ImageOrientationPatient with [1 0 0 0 1 0] ) which
> created new problems. I realize that I could use "Reorient images" on
> the display page, but prefer a more efficient solution.

This - or the "reorient images" plugin from the spm_orthviews plugins (see 
the SPM extensions page) - will not do the job either. What you need to do 
is "guess" new image voxel values, since you want to have data for 
coordinate points where you did not actually measure anything. 

This is the same question that is asked so often about reslicing images using a 
certain template. What you have to do:
- Create an image handle (see spm_vol for details) with suitable voxel 
size, image dimensions, transformation matrix etc (you will have to 
calculate these things in Matlab) and save an image with arbitrary data 
matrix using spm_write_vol (can be all zeroes, ones, random noise).
- Use realign/reslice only and select the image you just created as first, 
your MR image as second image. You will get a r*.img which has the proper 
orientation.

-- 
Volkmar Glauche
-
Department of Neurology		[log in to unmask]
Universitaetsklinikum Freiburg	Phone	49(0)761-270-5331
Breisacher Str. 64		Fax	49(0)761-270-5416
79106 Freiburg