An awesome library.
Dear Dirk, in file image_registration.m at line 533 to 543, shouldn't
come after computing filter size for Gaussian, i.e. after line 541
Since, O_trans is converted to vector, size(O_trans,2) is 1.
Thus, size(Istatic,2)/size(O_trans,2) = size(Istatic,2)
which is very large(around 130 for an image of size 400*530).