Neural Network Analysis for Prediction of Material Property in Human Eye with Keratoconic Cornea

Characterizing the material property of human eye isnecessary to provide any biomechanical solution of eyediseases. In this study, a new approach for prediction ofmaterial property in a human eye with keratoconic corneahas been presented by using Finite Element and NeuralNetwork analysis. In the Finite Element model, a hyperelasticcornea was used to obtain the deformation of corneadue to intraocular pressure. The amount of keratoconusextension was simulated by sectioning the cornea in toweakened angular section tied to a healthy section. TheFinite element model was used to obtain cornea curvaturefor different material properties. Obtained results were thenused as data points in the Neural Network program. Theregression and performance plots of implemented NeuralNetwork analysis suggest that predictions are reasonable andreliable. Trained network can be used for accuratelycharacterizing the material properties of keratoconic corneafrom curvature maps of keratometers.