Visual Acuity Prediction On Real-life Patient Data Using A Machine Learning Based Multistage System

This article discusses an IT system architecture that aggregates patient information from various ophthalmic data resources to make predictive statements about the progression of visual acuity in patients with certain eye diseases. The system uses deep neural networks and a patient progression visualization and modelling dashboard to aid ophthalmologists in making predictions about a patient’s visual acuity. The article also presents an evaluation formalism and compares the system’s predictions to those of human annotators.