Publication

Abstract : Retinal blood vessel segmentation is a preliminary step for screening of diabetic retinopathy and other retinal diseases. This paper proposes an automated system for segmenting the retinal blood vessels and thereby assisting the ophthalmologists for diagnosing diabetic retinopathy. The proposed system uses contourlet transform for the segmentation of vascular patterns. Contourlet transform is the combination of laplacian pyramid, directional filter bank and the wavelet. The evaluation of proposed system is done using the DRIVE database which contains 40 fundus images and its manually labelled images. In DRIVE database, the proposed system achieves an average accuracy of 94.67%. © Research India Publications.