Diabetic Retinopathy (DR) is one of the most common health problems of long-standing diabetes. It is a progressive disease and finally results in blindness. Early detection of diabetic retinopathy can prevent the damage to the retina and vision loss or at least slow its progression. The aim of the present study was to develop a Computer Aided Diagnostic System (CAD) in order to diagnose diabetic retinopathy by extracting blood vessels, optic disc and exudates from retinal digital fundus images and to validate the system with multiple image databases. The digital color fundus images obtained from 60 Indian women, 40 images from Digital Retinal Images for Vessel Extraction (DRIVE) database, 40 images from Structured Analysis of Retina (STARE) database and 89 images from DIARETDB1 database were used for the present study. For each image the hand drawn ‘ground truth’ result was collected. Then the quantitative analysis of the proposed algorithm with Support Vector Machine (SVM) classifier has been carried out and the results are compared against the ‘ground truth’. The extracted features such as mean intensity, mean area of the segmented region, number of segmented regions and solidity were displayed significant (p<0.001) differences between normal and diabetic retinopathy group for all test image dataset. The developed Computer-Aided Drafting (CAD) for diabetic retinopathy screening is validated with publicly available fundus image databases containing the ‘ground truth’ collected from several experts and comparison with private database. The proposed Support Vector Machine (SVM) classifier gives the sensitivity, specificity and accuracy values of around 92%. Hence, the proposed Computer-Aided Drafting (CAD) system could be useful for diabetic retinopathy screening.