Memristive Binary Firefly Algorithm for White Lung Images Segmentation

This paper investigates the feasibility of designing and employing swarm intelligence algorithms using memristors. As a representative swarm intelligence algorithm, the firefly algorithm is binarized and implemented through a memristor crossbar array to facilitate image threshold segmentation. Firstly, a memristor crossbar array with an external control circuit is meticulously designed, presenting both elevated device density and the potential for parallel computing. Secondly, the binary firefly algorithm is enacted through this memristor crossbar array. Finally, the method is put to practical use by applying it to the segmentation of medical images, the threshold can be directly calculated from the current and charge within the crossbar array. Notably, the error rate of our crossbar-based binary firefly algorithm implementation is significantly low compared to the brute-force method.