Artery Vein Classification of Blood Vessels in Retinal Image: an Automated Approach

Artery/Vein (A/V) of retinal vessel is helpful for the automating the detection of various diseases such as Diabetic Retinography, high blood pressure, pancreas, and other vascular conditions. An automated approach is presented in this paper for classification of blood vessels into Artery/Vein based on the features extracted from the centerline pixels. The proposed method consists of 4 important phases: 1) Segmentation 2) Centerline Extraction 3) Feature Extraction and 4) Classification. First input image is converted into green channel then retinal vessels are extracted using segmentation process after that centerline of the vessels is extracted using the thinning algorithm. Next step is to extract the features from the centerline pixels and finally vessels are classified using the extracted features. Input images are available on the databases like VICAVR [7], DRIVE[5] and INSPIRE-AVR[6]. Keywords-Retina, Vessel Segmentation, Centerline, A/V Classification