Daubechies wavelet‐based local feature descriptor for multimodal medical image registration

Abstract

A new local feature descriptor recursive Daubechies pattern (RDbW) is developed by defining andencoding the Daubechies wavelet decomposed center–neighbour pixel relationshipin the local texture. RDbW features are applied in spatial alignment(registration) of multimodal medical images using a Procrustes analysis(PA)-based affine transformation function and the registered images are furtherfused by employing a wavelet-based fusion method. A significant amount ofexperiments is conducted and the registration and fusion accuracy of theproposed feature descriptor is compared with the prominent existing methods suchas local binary patterns (LBP), local tetra pattern (LTrP), local diagonalextrema pattern (LDEP), and local diagonal Laplacian pattern (LDLP).Experimental results show the present registration method improves the averageregistration accuracy by 38, 47, 71, and 76% in contrast to LDLP, LDEP, LTrP,and LBP, respectively. Further, the fusion results of the current approachexhibit an average improvement in entropy by 11%, standard deviation by 6% edgestrength by 12%, sharpness by 23%, and average gradient by 16% when comparedwith all other feature descriptors used for registering the images. Conceptspresented here can be used widely in analysing the combined information presentin multimodal medical images.