Retinal Vessel Extraction and Vessel Path Prediction by Active Contouring

Diabetic retinopathy affects eye sight by damaging the small blood vessels present at the retina, vessels becomes narrow and micro aneurysms are formed inside the retina. In this proposed work, an efficient approach of filtering and thresholding is presented for the accurate extraction of retinal blood vessel. A systematic approach is used here for vessel extraction with less complex algorithm. Gaussian high pass filter is used for edge detection and global thresholding is done for contrast enhancement. To make the vessel structure continuous, Gabor filter is used.  For tracing the lost nerve path, Active contouring is performed on threshold image, which is an energy minimization process. An approach of convolution of Kirsch template with retinal image is used with high directional selectivity for the minute edges. This present system performs well for the extraction of vessels in retinal image