Neural Network Based Classifier for Retinal Blood Vessel Segmentation

A supervised method is proposed for automated segmentation of vessels in fundus images of retina. This method is used to detect the retinal diseases by extracting the retinal vasculature utilizing 9-D feature vector based on orientation analysis of gradient vector field, morphological transformation, line strength measures, and Gabor filter responses. The feature vector encodes information to handle the healthy and pathological retinal image. Each pixel in the retinal image is characterized by a vector in 9-D feature space and those pixels are classified using neural network classifiers (FFBNN. RBF, and MLP) and the performance is evaluated in detail. As its effectiveness and robustness with different image conditions, together with its simplicity and fast implementation, make this blood vessel segmentation proposal suitable for retinal image computational analyses such as automated screening for early retinal disease detection. KEYWORDSRetinal blood vessels, segmentation, neural classifiers, Feed Forward Back propagation Neural network (FFBNN), Radial Basis Function (RBF), Multi-Layer Perceptron (MLP), medical image analysis.