Diabetic Retinopathy (DR) is worldwide commonly observed disease related to eye. The number of persons who have gone blind or have damaged vision as a result of diabetic retinopathy has grown during the last 20 years. Due to diabetes the retina of eye has been affected. Diabetes also has an impact on the heart, the neurological system, and the kidneys. When a person has diabetes, their retinal blood vessels expand and begin to leak fluid and blood. Vision loss may occur if DR progresses to an advanced degree. DR is the main cause of blindness in the globe, accounting for 2.6 percent of all cases. Diabetics should have regular retinal examinations to detect and treat diabetic retinopathy (DR) early enough to avert blindness. In this post, we looked at the various approaches for diagnosing diabetic retinopathy. A group of lesions that appear in the retina of persons is called DR. Early detection of exudates may aid in the prevention of vision loss. This article discusses the approaches, algorithms, and methodologies used for pre-processing, segmentation, and detection of retinal images. Accuracy of various deep learning method is compared such as CNN, ResNet,DNN,Alexnet, InceptionNet, and VggNet.