Blood Vessel Extraction from Retinal Images Using Modified Gaussian Filter and Bottom-Hat Transformation

Abstract

Extraction of Retinal blood vessels are important for computer-aided diagnosis of ophthalmologic diseases. This paper presents a novel method for the detection of blood vessels of the retina. In proposed algorithm, at first, the background of the retinal vessels is removed. Then we perform modified Gaussian filtering to remove the noise of the image. The resulting image is subjected to Bottom-hat transformation and thereafter a width-dependant threshold operation is carried out to extract the blood vessels effectively. The proposed method is very efficient and less computational time as compared to the previously implemented methods and is thus very useful.