Survey on Machine Learning Techniques Used for Central Serous Retinopathy Detection

Central Serous Retinopathy (CSR) is an important health concern affecting millions of people worldwide, leading to vision loss and blindness. This condition is identified by abnormalities in retinal layers and fluid leakage, frequently occurring in the vicinity of the macula. The disease generates pressure inside the retinal layers, resulting in the detachment of retinal walls and subsequently causing impaired vision. As the macula is home to most of the visual sensory organs, including rods and cones, the eyesight undergoes substantial effects. Thus, early detection of CSR permits preventive actions to be taken to avoid any eye damage. While numerous physical approaches for detecting CSR have been devised previously, they have demonstrated inaccuracies and a lack of reliability. This paper examines the different algorithms, imaging modalities, and the outcomes they yield that have been devised over time for the timely detection of CSR. The research reveals that the SVM classifier has achieved a 97% accuracy rate in recognizing central serous retinopathy ailment based on OCT images.