An improved hybrid fusion of noisy medical images using differential evolution-based artificial rabbits optimization algorithm

This article investigates the problem of removing noise from multi-modal medical images to ensure efficient Medical Image Fusion (MIF). The proposed MIF achieves optimal results with a novel hybrid image fusion scheme. This scheme is achieved with an improved performance of the Artificial Rabbits Optimization (ARO) algorithm and a novel cascaded combination of filters. The exploring mechanism of the classical ARO algorithm is enriched by incorporating the approaches adopted in Differential Evolution and thus termed Differential Evolution-based Artificial Rabbits Optimization (DEARO). The effectiveness of the novel DEARO algorithm is proven through the testing of the CEC 2017 benchmark functions and it is noticed that the proposed approach offers superior solutions than existing optimization algorithms. Ten image fusion quality evaluation metrics are compared to demonstrate the performance of the proposed approach. Considering Mutual Information (MI), the proposed method exhibits \(40\%\) $40\%$ average improvements in the fusion of clean images. Similarly, \(50\%\) $50\%$ , \(36\%\) $36\%$ , and \(21\%\) $21\%$ improvements are noticed in MI values when both the modalities of source images are contaminated with Gaussian, Salt & Pepper, and Speckle noises of variance 0.1. The qualitative evaluation of the fused image shows the advancement of the proposed scheme in multi-modal MIF compared to the contemporary approaches.