AVR Estimation

OVERVIEW

This work is organized as follows. Chapter 4 focuses on design and implementation of the work system where we can diagnose the retinal image that normal or abnormal from the AVR that estimated. This encompasses the retinal vessel segmentation and the location of the vascular structures. It having various phases like vessel extraction, graph generation, A/V classification etc. are devoted to explain the most important phases in the AVR computation, these are, the measurement of the retinal vessel widths in the region of interest and their distinction in artery and vein types. AVR estimation is devoted to present two procedures for the AVR computation. The former computes the AVR automatically within a patient's image independently,

But for the processing these retinal images are first converted into Grayscale image, As we know the computational speed of it is more than color image. The Block Diagram for system are as follows. In some cases vessels of retinal image do not clear to see. Hence there is the need of enhancement. Here retinal image is enhanced by the Histogram Equalization method. Then the vessels of the retinal image are extracted. In the recent years graph have emerged as a unified representation and graph method have used in retinal image segmentation, retinal image classification and retinal image registration. Hence I used the approach of graph based classification. The work system uses the following step: Image Enhancement, Grayscale conversion, Image binarization, Morphological operation, Graph Generation, Feature extraction, A/V Classification and AVR

Here diameter is the important feature to classify the retinal blood vessels. These features are extracted by using the region properties for retinal image. Here automatic graph based approach is used for classifying retinal vessels into arteries and veins . The features gets extracted on the basis of centerline extracted image and a label is assigned to each centerline, indicating the artery and vein pixel. Based on these labelling phase, the final goal is now to assign one of the labels with the artery class (A), and the other with vein class (V). In order to allow the final classification between A/V [24] classes along with vessel intensity information the structural information and are also used. Classification is done with the help of SVM classifier. Here the important feature for the classification is the width of the vessels. With the help of SVM classifier we can easily seperate out the vessels into arteries and