With the increased number of people suffering from diabetes, there is an urgency to automate Diabetic Retinopathy Detection. According to the International Diabetic Federation, there are over 93 millions of patients suffering from Diabetic Retinopathy world-wide. Currently, there are three types of manual screening procedures to detect Diabetic Retinopathy which include Pupil Dilation, Ophthalmoscopy and Tonometry. Several attempts to automate the screening of Diabetic Retinopathy has been made over the last two decades. An Automatic Diabetic Retinopathy Screening System (ADRSS) serves as a supporting tool for an Ophthalmologist to increase the speed and accuracy of Diabetic Retinopathy Detection. In the domain of research, several such systems have been made available which makes use of ocular images different imaging modalities. Depending upon different modalities, different image processing algorithms have been put in practice. In this juncture, this proposed aims to study and analyze all the different detection techniques available with an emphasis to techniques for respective image modalities. This proposed paper will conduct a detailed review on the different algorithms that are being implemented of the different types of images from different modalities with their diagnostic level of accuracy in Diabetic Retinopathy Screening.