When initializing an affine transform for a registration framework, what is the difference between the offset and the fixed center inputs?
Please see this jupyter notebook for all the details with respect to spatial transformation.
I think i am just confused about initializing. Because when I use an exhaustive search without initialization, the images register incorrectly but biased in the top left corner (image origin).
So I am wondering how I can get this to either a) actually search the full image, or b) initialize it closer to the answer using known locations?