Developing Automated Bifurcation Angle Measurement Algorithm for Retinal Image Analysis

Retina and its vascular network have unique branching characteristics morphology. The abnormalities in the vascular pattern of the retina, such as vessel shape, branching pattern, width, tortuosity, or the appearance of retinal itself, may be associated with the occurrence of retinopathy or cardiovascular diseases. This paper is focus on the development of an algorithm to measure bifurcation angle of the retinal for image analysis. From survey, there is still lack of automation for bifurcation angle measurement. Thus this initial work investigates and proposes the key steps for bifurcation measurement process. The key steps of the proposed method include segmentation of the region of interest (ROI), tracing the vessels that are needed for angle determination and lastly determine the bifurcation angle by using the dot product of the traced boundaries. The results are compared with the manual measurement of the bifurcation angle using quantitative and qualitative measurement. The experimental results show promising result of the proposed method in automating the measurement of bifurcation angle of retinal. Recommendation for room of improvement is also discussed in the paper.