Retinal Vasculature Extraction using Wavelets

Abstract

Retinal vasculature extraction helps in diagnosing the early detection of diabetic retinopathy to prevent blindness. In this work vasculature structure extraction is proposed which is based on vessel detection method. Retinal normal and abnormal images are first preprocessed to enhance the vessel information. Two methods i.e. wavelet transforms along with multiscale hessian-eigenvalue approach are considered for vessel extraction. Results are promising and it shows that the method performs well in extracting the vascular pattern. The comparative studies with manual based segmentation prove the effectiveness of our proposed method.