This is likely because your data is in neurological display convention
(check using fslorient <data>). If you store your data in radiological
convention you won't have this problem. Some conversion softwares, e.g
dinifti unfortunately default to storing the data neurologically, even if
that was not originally the correct orientation.
Peace,
Matt.
-----Original Message-----
From: FSL - FMRIB's Software Library [mailto:[log in to unmask]] On Behalf
Of Corina Melzer
Sent: Thursday, March 13, 2008 5:53 AM
To: [log in to unmask]
Subject: [FSL] FAST - L-R-Orientation
Hi,
I have some troubles when using the FAST segmentation algorithm. If I
overlay the original image
and one of the resulting masks there is a left-right orientation mismatch
between these two images.
Checking the header information (fslhd) doesn't show any orientation
differences between the
images. They're both anatomical and the orientation of x-y-z is the same
(L-R,P-A,I-S).
Fslview shows the correct orientation labels for the original image and I am
sure, that it's orientation
is ok. Only the resulting segmentation mask after FAST is wrong.
At the moment I'm handling this problem by reorienting the mask using
fslswapdim before masking
the original image with fslmaths. But in my opinion this is not an optimal
solution.
Maybe you have an idea, where the problem could be.
Thanks,
Corina
|