A wise colleague of mine once likened judging registrations to what the US supreme court said about art vs. ... explicit images which are not art: It may be hard to define but you know it when you see it.
I think the best practice is to look at it and make a judgement. Overlay the image on the template and see if there is any regions where it is a bad fit. The top of the brain should be smooth when you look at a sagittal view, not all bumpy and distorted (though check the unregistered MRI for unusual shape as well). Look at several scans together: if they are well normalized, they should overlap quite nicely. If the brains do not have a fairly consistent shape or one has a different shape/appearance than the other, this suggests a problem.
Good luck!
Colin Hawco, PhD
Neuranalysis Consulting
Neuroimaging analysis and consultation
www.neuranalysis.com
[log in to unmask]
-----Original Message-----
From: FSL - FMRIB's Software Library [mailto:[log in to unmask]] On Behalf Of Sam
Sent: March-16-16 5:13 AM
To: [log in to unmask]
Subject: [FSL] Nonlinear Registration Quality
Dear Experts,
I would like to understand more about what I should be looking for to help me know whether registration should be considered 'good' - particularly nonlinear. I have looked at the Wiki, FSL course and mailing list, but not found anywhere with any detail (ideally with pictures). Can anyone direct me to any resources please?
Thanks.
Sam.
|