Segmentation of Blood Vessels from Retinal Fundus Images Using Bird Swarm Algorithm and River Formation Dynamics Algorithm

Abstract

In this work, we have presented a new bio-inspired blood vessel segmentation hybrid technique using bird swarm algorithm (BSA) and river formation dynamics (RFD) algorithm. BSA mimics the behavior of the birds and RFD imitates the process of formation of riverbeds. The birds pass through each and every pixel while flying and foraging for the food in order to identify the edge pixels falling in their path. The path taken by the birds has been optimized with the help of RFD algorithm. The pixels are categorized into the edge or non-edge pixels by thresholding. The proposed technique has been implemented and evaluated quantitatively and qualitatively on STARE dataset for the analysis. The experimental results clearly showcase the remarkable improvement over a few metaheuristics and non-metaheuristic techniques namely flower pollination search algorithm, ant colony optimization algorithm and matched filter in terms of accuracy, specificity, and sensitivity. The defined approach gives promising results by detecting continuous and smooth blood vessels with an accuracy of 95.05%.