Extraction and Classification of Blood Vessel Minutiae in the Image of a Diseased Human Retina

摘要

The work presents a created methodology for detecting minutiae in the image of the retina of a sick person's eye. The worked out algorithm is used to find areas of distribution and classification of minutiae in ill human eye retinal image. The main goal of the proposed approach is to classify all minutiae into groups based on the distance from the center of the image and the distance from the edge of the image that is closer to the blind spot. For the separation of blood vessels from the image, Otsu algorithm and background subtraction were used. To get line representation of blood vessels that helps find minutiae in images, the K3M thinning algorithm was used. The proposed algorithm shows one of the most basic solutions for finding the characteristic points in a biometric image. The last step of the presented algorithm introduces an example of minutiae classification.