Hi all,
I am trying to understand a bit more about Dartel and the qform matrix.
I have acquired structural images from 4 timepoints. I used the
longitudinal registration toolbox to warp all timepoints to the half-way
space and then segmented all 4 timepoints separately. The qform matrix
(under V.private.mat0) for the imported rc* images is slightly different
for each timepoint, although the timepoints are in exactly the same
space and the qform matrix for the c* images is identical.
Since the original images are all in exactly the same space and even the
c1/c2 images have identical information across all timepoints, I'm not
sure why there would is a difference in the qform matrix for the
imported images. The differences are nevertheless small, but I'm not
sure if they are important.
Below is an example for two imported images (flimg and flimg1):
flimg.private.mat0
ans =
-1.4999 -0.0097 0.0076 89.9203
-0.0122 1.3000 -0.7482 -61.0120
0.0018 0.7482 1.3001 -115.9505
0 0 0 1.0000
flimg1.private.mat0
ans =
-1.4999 -0.0089 0.0098 89.6944
-0.0125 1.2992 -0.7496 -60.7169
0.0040 0.7496 1.2992 -116.1226
0 0 0 1.0000
The qform and sform matrices for the original images (which were
segmented to produce the imported ones) are exactly the same in both cases:
flimg0.private.mat0 and flimg0.private.mat
ans =
0.7991 0.0024 -0.0385 -97.6872
-0.0236 0.6619 -0.4487 -9.6544
0.0305 0.4493 0.6612 -218.7335
0 0 0 1.0000
Many thanks in advance for the help,
Marina
|