Advances in Computer-Aided Detection and Diagnosis of Retinal Diseases: A Comprehensive Survey of Fundal Image Analysis

Diseases pertaining to the Retina are the leading causes of blindness comprising globally 22–25% of people suffering from it. The go to first line of diagnosis for ophthalmology processes to detect retinal diseases comprise of fundoscopy in which the fundus of image is dilated and subsequent imaging is performed in order to identify the probable diseases and predict possible outputs. The paper is a survey of latest accomplishment made in the defied in CADe-CADx domain with respect to fundal images and their associated diseases. It encompasses various image processing and classification techniques proposed by researchers in recent times. The current research scenario covers a variety of topics ranging from traditional CNNs to modern techniques like Visual Transformers. It explores multiple models used for classifying different numbers of diseases. Various performance metrics as used by various researchers are also discussed.