3D Gabor Wavelets for Evaluating Medical Image Registration Algorithms

摘要

A Gabor wavelets based method is proposed in this paper for evaluating and tuning the parameters of image registration algorithms. The registration quality is measured by the anatomical variability of the registered images. We propose in this paper a local anatomical structure descriptor, namely the Maximum Responded Gabor Wavelet (MRGW) for such a purpose. The effectiveness of the descriptor is demonstrated through a practical spatial normalization example - the variance of MRGW is successfully applied to tune the parameters of a nonlinear spatial normalization algorithm, which is integrated in one of the most popular software packages for medical image processing - the Statistical Parametric Mapping (SPM).