Swarm intelligence techniques in segmenting human retinal vasculature

Abstract

Vasculature of human retina furnishes information concerning various eye related ailments and also assists in lesions detection. Severity of the eye diseases can be discerned from pathological conditions related to changes in the retinal vasculature. In this work, for pre-processing, Contrast Limited Adaptive Histogram Equalization (CLAHE) and average filter is used to enhance the input image. Further, swarm intelligence techniques, Particle Swarm Optimization (PSO), Darwinian Particle Swarm Optimization (DPSO), Fractional Order-Darwinian Particle Swarm Optimization (FO-DPSO) are used in segmenting the blood vessels of the human retina. Additionally, similarity index metrics are employed in evaluating the accuracy of the retinal vasculature segmentation with ground truth. The results obtained clearly reveals that FO-DPSO outperforms in segmenting accurately than PSO and DPSO. Results of the segmentation are further reinforced using box and dendogram plot.