Ratanapakorn T, et al. - To generate automated software for screening and diagnosing diabetic retinopathy (DR) from fundus photographs of subjects with diabetes mellitus, researchers evaluated a set of 400 fundus images, containing 21 normal fundus images and 379 DR fundus images including 162 non-proliferative DR and 217 proliferative DR. Initially, they observed 98%, 67% and 96.25%, the sensitivity, specificity and accuracy of software in detection of DR, respectively when compared to that of the diagnosis by the ophthalmologist. Although, they found 66.58% of the accuracy of this software in the classification of non-proliferative and proliferative diabetic retinopathy. They suggested it a helpful tool for DR screening in the distant rural regions where any ophthalmologist was not available.