Diabetic Retinopathy Classification using Deep Learning Algorithm

Image processing is an important and challenging factor in the field of medical sciences. It is widely used for the disease detection, classification and to improve the diagnosis process. This paper deals with fundus image disease analysis which is caused by diabetics that is called diabetic retinopathy image analysis. Here the deep convolution neural network model is helps to analysis the fundus image for improving diagnosis process. Major steps to analyze the eye disease image is enhancement, filtration, segmentation, and classification. These are all the basic steps to improve the diabetic retinopathy diagnosis treatment. In this advanced diabetic eye disease analysis, deep learning technique is used majorly for segmentation and classification. Image morphological process is performed initially for image quality enhancement and filtration. After that deep learning technique is performed with convolution layers to classify the image. This proposed model is performed to improve the accuracy of results when compared to previous literatures. In results the geometric transformation and set of accuracy, sensitivity and specificity is calculated.