Feature Analysis for Segmentation of Retinal Blood Vessels

Abstract

Retinal Vascular Segmentation proves useful for clinical purposes. From segmented blood vessel, tortuosity and diameter of vessels can be computed. Vessel tortuosity can be used for evaluation of hypertensive retinopathy whereas vessel diameter can be used to diagnose hypertension and cardiovascular diseases. In this paper, we have extracted the features for the detection of blood vessel. The extracted feature is vesselness measures, matched filter, Gabor filter and matched filter. Later these features are combined using multifeature analysis with the help of neural network for the segmentation of retinal blood vessels. Key Words: Inverted green channel, Retinal Vascular Diseases, Retinopathy.