New techniques for segmentation and extraction retinal blood vessels

Abstract

Blood vessels are one of the important organs in the retina that can be used to diagnose various diseases. We research blood vessel segmentation and extraction with various methods in the STARE and DRIVE dataset. The methods we use are adaptive thresholding, otsu thresholding, isodata, fuzzy c means, and bradley threshold. The parameters we use as a result of comparison are accuracy, sensitivity, and specificity. The best results of accuracy, sensitivity, and specificity in the STARE dataset were 93.66%, 92.62%, and 96.85%, respectively. The best results of the DRIVE dataset are 95.96%, 90.37%, and 98.36%, respectively.