Classification of Eye Diseases and Detection of Cataract using Digital Fundus Imaging (DFI) and Inception-V4 Deep Learning Model

Ophthalmologists use retinal imaging to diagnose a variety of eye disorders, including microvascular retinal disease, which arises in consequence of high blood pressure and diabetes. Periodic ophthalmoscopy is the most effective method of screening for eye diseases. However, the scarcity of ophthalmologists is a barrier to beginning inspection. The existence of digital fundus cameras for automated image processing can assist ophthalmologists in diagnosing eye illness. Cataracts, glaucoma, and other retinal diseases are the most prevalent causes of age-related eye disorders and vision deterioration in the elderly. The adoption of a computer-based intelligent approach for the categorization of various eye disorders is extremely beneficial in both diagnosis and disease prevention. This study describes a deep learning-based categorization approach for four types of digital retinal images (DRI). Invariant of Inception v4 model is tested on a Kaggle database of 602 DRI of 1.67 GB. We achieve a 96% accuracy rate, and the findings are extremely encouraging.