Use of steerable wavelets in medical image registration
Sonal Gahankari, Poornima Talwai, Bhavesh Parmar · 2009
Image registration has been widely used in many works of image processing. In this paper, a novel idea of medical image registration using steerable wavelets is proposed. Multiresolution image registration approach has potential to speed up the computation. Here, rotation and translation invariant steerable wavelets are used for multiscale and multi-orientation image representation, to overcome the limitations of conventional wavelets. Registration results of coarsest resolution scale are directly passed to the finest resolution scale to speed up registration. Registration similarity measure of mutual information is used in the coarsest scale and normalized cross correlation is used in the finest scale, the experiments on medical datasets demonstrates efficiency of the proposed algorithm. (4 pages)