Hi Mark,
I have major registration issues!
I am consistenly getting registration errors between my T2*s and SPGRs. Typically
it's a scaling issue along the z-axis. I've tried a variety of DOFs and most of the cost
functions. Sometimes it gets close, but it's always visibly misregistered.
I then tried to improve my registration using fugue. Now, the problem is that I can't
register the T2 to the SPGR which means I can't apply the phasemap to the SPGR.
The T2 is 256x256x16 and is fairly high contrast. The SPGR is 256x256x124 and is
even higher contrast. The FOVs are the same in plane but not the same along the z-
axis. So this is the crux of the problem, right? You could maybe pare down the
SPGR until its span along z is approximately the same as the T2, but that's not very
precise and defeats the whole purpose of automated registration. There appears to
be enough contrast in the images such that flirt should find the global minimum. But,
alas, no.
So, in order to improve the registration of my T2*, I must first register the T2 to the
SPGR, and since I can't do either, I 've turned into a disillusioned and bitter shell of
my former self!
jack
|