PROJECT TITLE :

An Automatic Graph Based Approach For Artery Vein Classification In Retinal Images - 2014

ABSTRACT:

The classification of retinal vessels into artery/vein (A/V) is an important part for automating the detection of vascular changes, and for the calculation of characteristic signs related to many systemic diseases such as diabetes, hypertension, and different cardiovascular conditions. This paper presents an automatic approach for A/V classification primarily based on the analysis of a graph extracted from the retinal vasculature. The proposed methodology classifies the complete vascular tree picking the sort of every intersection point (graph nodes) and assigning one of two labels to each vessel phase (graph links). Final classification of a vessel section as A/V is performed through the mix of the graph-based labeling results with a collection of intensity features. The results of this proposed method are compared with manual labeling for 3 public databases. Accuracy values of 88.3percent, eighty seven.4p.c, and 89.eightp.c are obtained for the photographs of the INSPIRE-AVR, DRIVE, and VICAVR databases, respectively. These results demonstrate that our method outperforms recent approaches for A/V classification.