Glow Worm Optimization based ANFIS With Mahalanobis Distance for Effective True Blood Vessel Detection

Abstract

Automated retinal analysis progresses primary recognition and exploration of diseases like diabetic
retinopathy. To increase the retinal diagnosis results, the blood vessel segmentation is intricate in most
prevailing researches. Conversely the prevailing approaches have noise and exactness concerns. In this paper,
Glow worm optimization based ANFIS with Mahalanobis Distance (GWO-ANFIS-MD). Primarily, the image
denoising is completed by using Non Local Means (NLM) filter followed by Adaptive Histogram Equalization
(AHE) for image development. The appropriate and important structures are removed by using Modified
Particle Swarm Optimization (MPSO) with Tabu Search (TS) in retinal images. Then GWO based ANFIS is
used to implement the vessel detection more successfully in this research. The investigational result evidences
that the vessel detection presentation is greater in terms of advanced accuracy, sensitivity, specificity and fmeasure
by using the proposed GWO-ANFIS-MD method.