• Poonam Namdev Shelke

Retina is light-sensing tissue that resides in the back of the human eye. Retina is the one which relays the image to brain. Visual impairment and blindness is caused due to many retinal diseases such as Glocoma, Macular Degeneration(MD) and Diabetic Retinopathy(DR). The patients having DR is more than patients having other retinal diseases. Expert ophthalmologists are required to detect DR. Even these experts need to spend lot of time to screen the images and diagnose the disease. Thus automated system can help ophthalmologists not only save screening time but also to diagnose the disease with more accuracy. Proposed work is aimed at automating DR detection using deep learning algorithms. Convolution Neural Network(CNN) is used for this purpose. To detect DR, it may not be effective with few number of features and as CNN can create more number of features leading to accurate prediction. In this proposed work, Different CNN architectures are used to detect diabetic retinopathy. Proposed work is to prepare input images for CNN using image preprocessing techniques such as image contrast enhancement, Median filter. Total 516 images are used for training and testing. CNN with 16 layer architecture is used for classification.The performance of system gives 60% of accuracy on IDRiD dataset.

CNN, Deep Learning, Diabetic Retinopathy, Image Processing, Retinal Disease.

"Diagnosis of Diabetic Retinopathy using Deep Learning", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.6, Issue 6, page no.296-301, June 2019, Available :http://www.jetir.org/papers/JETIR1907F39.pdf

"Diagnosis of Diabetic Retinopathy using Deep Learning", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.6, Issue 6, page no. pp296-301, June 2019, Available at : http://www.jetir.org/papers/JETIR1907F39.pdf