So, I have to retract this question; looking at a pial surface it seems pretty clear that the image on the right is a sensible fiducial surface for my data. That's just the anatomy.
But the the question becomes: in landmark registration, where a feature is missing in one surface and present in another, is it best to draw a landmark where the sulcus *should* be on the surface where it is absent (as in figure), or just remove that border from both surfaces. I would have guessed the latter. I suppose one can try both. On 12 July 2011 20:31, <[email protected]> wrote: > Send caret-users mailing list submissions to > [email protected] > > To subscribe or unsubscribe via the World Wide Web, visit > http://brainvis.wustl.edu/mailman/listinfo/caret-users > or, via email, send a message with subject or body 'help' to > [email protected] > > You can reach the person managing the list at > [email protected] > > When replying, please edit your Subject line so it is more specific > than "Re: Contents of caret-users digest..." > > Today's Topics: > > 1. segmentation for surf morphometry issue (Colin Reveley) > > > ---------- Forwarded message ---------- > From: Colin Reveley <[email protected]> > To: <[email protected]> > Date: Tue, 12 Jul 2011 20:10:37 +0100 > Subject: [caret-users] segmentation for surf morphometry issue > Hello - > > I'm having a segmentation difficulty. The aim is to register F99 (macaque) > against an individual. The individual was scanned ex-vivo, so it's an MTR > image rather than T1. So the contrast is kind of better than T1 and that > may, or may not cause issues with surefit. I've tried segmentation many > ways. > > the data is 0.25mm, so it is is downsampled to 0.5mm. there is a bias field > but I think it's corrected pretty well (using FAST, in FSL) and the > intensities are uniform. > > The MTR image has to be put into the correct orientation for caret, so I > register it to a deskulled F99. I've tried 6 parameter (I see no need to > scale it) and affine transforms. I transorm at 0.25mm (I upsampled the > deskulled F99), and then downsample for surefit. > > whatever I try, the result is the same as the figure attached, on the right > next to F99. > > I can tolerate most of the differences, but the issue at the rostral end, > where there is a fork in the frontal lobe towards the posterior that is > simply absent in F99, F6, wisconsin and all macaque literature. It worries > me a lot. It seems a lot of variation for macaque. > > In the left hemisphere, this problem is less pronounced, but I'd like to > use the right for obvious reasons. > > It seems like this is a difference in gross anatomy. So it's likely wrong. > > when looking at the surefit result, there is quite a bit of gray matter > that, it would seem, has no associated white matter, and hence is rejected > from the segmentation. > > I think it does have associated white matter (it must do), and the problem > is to do with the linear transforms and/or resampling. Or the MTR, just > maybe although I doubt it. > > In any case, I'd be grateful if you'd take a look. the region I'm referring > too should be clear from the figure. There's a missing sulcus from my brain, > and one that shouldn't be there, in the frontal lobe above the temporal tip. > > help appreciated. > > Colin, > > Sackler Centre > Sussex > > _______________________________________________ > caret-users mailing list > [email protected] > http://brainvis.wustl.edu/mailman/listinfo/caret-users > >
_______________________________________________ caret-users mailing list [email protected] http://brainvis.wustl.edu/mailman/listinfo/caret-users
