Today DRTS is necessary for diabetic retinopathy (DR) screening as an alternative to retinopathy screening by an ophthalmologist or optometrist.This paper presents an automated DRTS for the first time in Iran. By using this system patient could be screened without referral to specialist centers. Therefore, it will decrease the cost of transportation of patients, and will save time. In addition, the system can automatically screen retinal images of diabetic patients. So part of the work of ophthalmologist and optometrist can be done automatically by this system.

DRTS process involves: 1) Taking digital images of the retina of diabetic patients in the primary care physician's office, and electronically transmitting these images to a reading center for evaluation for DR. 2) The system automatically evaluates the quality of the received images. If the quality of the images is poor then a message will be sent to the photographer to re-capture images with better quality. 3) Good quality images will be assigned to readers in the system. In this level reader can be a non-physician trained person or Automated Image Processing Robot (AIPR) that his job is to separate the normal and abnormal images from each other. 4) After separation, abnormal images (images that are suspected to be the symptoms of DR) are sent to a specialist for diagnosis of diabetic retinopathy. In this level the specialist can be an ophthalmologist or AIPR that his job is to determine the final status of patients for referral and non-referral to specialist centers for diabetes.

This paper introduces a DRTS that can automatically evaluate the quality assessment of retinal images and diagnose DR in retinal patient images. The DR detection algorithm was tested on Messidor database and accuracy is 0.74, sensitivity is 0.77 and specificity is 0.71. Of course we are trying to improve the algorithm in the future so that we can achieve better results.

This Diabetic Retinopathy Telescreening system with the ability for automated separation of normal and abnormal images and automated detection of presence of DR will be a great help in screening of DR in the future.