A Deep Learning Ensemble Method to Visual Acuity Measurement Using Fundus Images

Visual acuity (VA) is a measure of the ability to distinguish shapes and details objects at given distance spatial resolution visual system. Vision one basic health indicators closely related person’s quality life. It first tests done when an eye disease develops. VA usually measured by using Snellen chart or E-chart from specific distance. However, in some cases, such as unconsciousness patients diseases, i.e., dementia, it can be impossible traditional chart-based methodologies. This paper provides machine learning-based measurement methodology that determines only based on fundus images. In particular, levels VA, conventionally divided into 11 levels, are grouped four classes three learning algorithms, SVM model two CNN models, combined ensemble method order predict corresponding level image. Based performance evaluation conducted randomly selected 4000 images, we confirm our estimate with 82.4% average accuracy for which each class Class 1 4 identifies 88.5%, 58.8%, 88%, 94.3%, respectively. To best knowledge, this measurements images deep learning.