DeepRetino: Ophthalmic Disease Classification from Retinal Images using Deep Learning
Fatima Zahra Belharar, Nabila Zrira
Abstract
Eye diseases are one of the main causes of visual impairment. Their causes are various: they may be related to the aging process or originate from another pathology, such as complications of diabetes. Therefore, early diagnosis is highly recommended to prevent and control eye diseases. Previous approaches focused only on the detection of glaucoma, cataract or diabetic retinopathy. The main purpose of this article is to propose DeepRetino, an automatic multi-classification approach for six eye diseases based on advances in Deep Learning, in particular Convolutional Neural Networks (CNNs). In the preprocessing phase, we first focused on the histogram equalization method called Contrast Limited Adaptive Histogram Equalization (CLAHE) to improve the contrast of the fundus images. On the other hand, in the learning phase, we initialize and update the network weights using Xavier Orthogonal and Adam Optimizer. Finally, we evaluate DeepRetino on the Ocular Disease Intelligent Recognition (ODIR) dataset for deployment.