Abstract
High computational cost is the main obstacle to adapting globally optimal branch-and-bound algorithms to intensity-based registration.Eristing techniques to speed up su ch algorithms use a multiresolution pyramid of images and bounds on the target functi on among dif ferent resolutions for rigidly aligning two simages.In this paper,we propose a dual algorithm in which the optimization is done in the Fourier domain,and multi-ple resoluti on levels are replaced by multiple frequency bands.The algorithm starts by computing the target function in lower frequency bands and keeps adding higher frequency bands until the current su bregion is ei-ther rejected or divided into smaller areas in a branch and bound manner.Unlike spati al multiresolution ap-proaches,to compute the target function for a wider frequency area,one just needs to compute the target in the residual bands.Therefore,if an area is to be dis-carded,st per forms just enough computations required for the rejection.This property also enables us to use a rather large number of frequency bands compared to the limited number of resoluti on levels used in the space dom ain algorithm.Erperimental results on real im ages demonstrate consi derable speed gains over the space do-main method sn most cases.