Vessel extraction from retinal images by using matched filter and first derivative of Gaussian function

Abstract: Accurate detection of blood vessels from retina is an important job in computer aided diagnosis of Diabetic Retinopathy. The Matched filter is simple & effectively good method for vessel extraction. But Problem with that it recognizes vessels and non-vessels edges so it will detect false vessels. In this paper we introduce new extension of Matched Filter i.e. FDOG to extract the vessels. The MF-FDOG is combination of Matched filter which is zero mean Gaussian Function and First order derivative of Gaussian. The vessels are detected by thresholding the response of Matched Filter and Threshold is adjusted by image response of FDOG. The proposed method is very simple and it reduces false detection of Matched filter and detects all fine vessels which are not detected by matched filter. It gives good competitive results of vessel detection with low complexity.

Keyword: Diabetic Retinopathy (DR), Matched filter (MF), First order derivative of Gaussian Function (FDOG).