An Effective Threshold Based Technique for Retinal Image Blood Vessel Segmentation on Fundus Image Using Average and Gaussian Filters

The fundamental components of automated retinal blood vessel segmentation for eye disease screening systems are segmentation algorithms, retinal blood vessel datasets, classification algorithms, performance measure parameters and noise removing filters etc. This paper is presenting retinal blood vessel segmentation techniques on 2D retinal images captured by fundus camera and examining the performance of different algorithms for a better and accurate result. The goal of this study is to comprehensively investigate and assess retinal blood vessel extraction algorithms using publicly accessible High-Resolution Fundus Diabetic Retinopathy (HRF-DR) and Iowa Normative Set for Processing Images of the retina-artery vein ratio (INSPIRE-AVR) datasets, with a focus on important performance measurements and characteristics. The effectiveness of the proposed methodology is measured using the performance metrics like Sensitivity, Specificity and Accuracy with 94.27%, 87.86% and 95.61% for HRF-DR dataset and 91.10%, 86.97% and 95.16% for INSPIRE-AVR dataset respectively.