Hi John,
I read from the spatial normalization help that " ... The matching involved
simultaneously minimizing the bending energies of the deformation fields and
the residual squared difference between the images and template(s). ..."
My question is that do you either 1 )globally scale the pixel values of the
source and target images to a comparable level before normalizing, or 2)
leave the scaling factor as one of the free parameters, or 3)... My concern
is that otherwise the "residual squared difference" would be a problem if
the pixel value ranges of the source and target images are too different.
Thanks for your explanation,
James Yu
|