A Comparative Study on Feature Selection for Retinal Vessel Segmentation Using Ant Colony System

The diabetic retinopathy disease spreads diabetes on the retina vessels thus they lose blood supply that causes blindness in short time, so early detection of diabetes prevents blindness in more than 50% of cases. The early detection can be achieved by automatic segmentation of retinal blood vessels which is two-class classification problem. Features selection is an essential step in successful data classification since it reduces the data dimensionality by removing redundant features, thus minimizing the classification complexity, time and maximizes its accuracy. In this paper, comparative study on four features selection heuristics is performed to select the best relevant features set from features vector consists of fourteen features that are computed for each pixel in the field of view of retinal image in the DRIVE database. The comparison is assessed in terms of sensitivity, specificity and accuracy of the recommended features set by each heuristic when used with the ant colony system algorithm. The results indicated that the recommended features set by the relief heuristic gives the best performance with sensitivity of 75.84%, specificity of 93.88% and accuracy of 91.55%.