Medical Image Registration Based on Simplified Multi Wavelet Transform

Abstract

Medical images registration technique is important to the modern clinical medicine. A registration algorithm of medical images is proposed in the paper. The algorithm consists of two part, i.e. the rough registration and the fine registration. In the rough registration procedure, the contour lines of the couple medical images are obtained at first, then they are resampled to obtain fewer contour feature points. The rough registration is accomplished based on the resampled contour lines and the Principal Axes method. In the fine registration procedure, the simplified multiwavelet transform is used to obtain the couple images multiscale information at first, then part of high-frequency coefficients are selected as registration objects. In addition, the selected high-frequency sub-bands are filtered by their mathematical expectation, and hierarchical registration strategy is used in the registration process. The experiment results show that the proposed registration algorithm is an effect method, not only the registrations precision can be held, but also the calculation cost is reduced.