Retinal Image Processing and Classification Using Convolutional Neural Networks

This study aims to develop a system to distinguish retinal disease from fundus images. Precise and programmed analysis of retinal images has been considered as an effective way for the determination of retinal diseases such as diabetic retinopathy, hypertension, arteriosclerosis, etc. In this work, we extracted different retinal features such as blood vessels, optic disc and lesions and then applied convolutional neural network based models for the detection of multiple retinal diseases with fundus photographs involved in structured analysis of the retina (STARE) database. Augmentation techniques like translations and rotations are done for expanding the number of images. The blood vessel extraction is done with the help of morphological operations like dilation and erosion and enhancement operations like CLAHE and AHE. The optic disc is localized by the methods such as opening, closing, Canny’s edge detection and finally thresholding the image after filling the holes. The bright lesions (exudates) inside the retina are detected by the filtering operations and contrast enhancement after the removal of the optic disc. In this study, we experimented with different retinal features as input to convolutional neural networks for effective classification of retinal images.