Unimodal Medical Image Registration Using Elite Opposition Bacterial Foraging Optimization Algorithm

Abstract

Medical imaging applications frequently use image registration for a variety of purposes, and the search of an ideal image transformation parameters that align the two images (reference and floating) is still an optimization challenge. Medical image registration has been optimized using different metaheuristics optimization strategies. One method, the Bacterial Foraging Algorithm (BFOA), has issues of poor exploration and low convergence to a better solution. This research work presents the Elite Opposition Bacterial Foraging Optimization Algorithm (EOBFOA) for optimizing unimodal medical image registration. The EOBFOA is an enhanced version of Bacterial Foraging Algorithm (BFOA) using the Elite Opposition Strategy. The proposed EOBFOA uses Root Mean Square Error (RMSE) as a measure to determine the accuracy of the image registration process. The performance of the image registration using the EOBFOA was compared against other existing nature inspired algorithms. The obtained results shown that the proposed EOBFOA outperformed other algorithms in searching for the best optimum transformation parameters for the image registration.