DarkNet-19 based Decision Algorithm for the Diagnosis of Ophthalmic Disorders

Performing different medical examinations primarily of which are visual examinations can prove to be beneficial for identification of retinal diseases. An upcoming interdisciplinary technology; computer-aided medical diagnostic system eliminates one of the key issues in the visual diagnosis of diseases caused by human error due to poor decision making. It provides precise detection and prediction of disease. In our proposed methodology, images were treated with Contrast Limited Adaptive Histogram Equalization followed by deep features extraction from the enhanced images. ADASYN was also applied to acquire a balanced dataset. Support Vector Machine classifier has yielded a mean accuracy of 93.8%.