> My first question is: Why is the origin off, or how was the
> origin determined?
If data are (9 parameter) affine registered with an unusually shaped brain,
then it is not possible to get all the landmarks in perfect agreement.
Something has to give somewhere. The construction of the original average
involved registering a bunch of brains by landmark matching. The idea was to
get the brain images inso positions so that the landmarks were as close
together as possible. If the position of the AC was made exact, then the
positions of the other points would have been more wrong.
Best regards,
-John
|