Hi,
it is correct although most people tend to register their diffusion data / results to structural space (as you typically want to take advantage of its higher resolution, despite all possible difficulties).
Anyway, your applyxfm-procedure is correct as well and the "expansion" of the mask is likely to be an interpolation and thresholding / display effect. Note that if you really want to go from str2diff you may as well try nearest neighbour for the mask. Check "flirt" for command line help.
Cheers-
Andreas
________________________________
Von: FSL - FMRIB's Software Library im Auftrag von Live Eikenes
Gesendet: Do 15.11.2007 12:33
An: [log in to unmask]
Betreff: [FSL] Registration in Fdt
Dear FSL users,
I want to use probtrackX, but first I have to register my structural
image to the diffusion space by using the registration after running
Bedpost. I wanted to check how the registration went, and used the
ApplyXFM GUI to apply the transform matrix on my structural image:
Transformation Matrix: str2diff.mat
Input Volume: the structural image, and the mask (or the ROI) that I
made in the structural image
Output SIze: Base on: Existing Volume: I chose here my nodif_brain image
as the reference volume
Is this the correct way to do it?
Escpecially the last part, is it correct to use the nodif_brain image as
the reference volume.
When I check the results, the registration of the structural image to
the diffusion space is ok, but the mask (or the ROI) selected in the
structural image, becomes much larger in the diffusion space.
Do someone have any idea what might have happened?
Where did i do wrong?
Hope to get some answers!
Best regards, Live Eikenes.
--
|