Machine Learning-Based Glaucoma Detection Using Fundus Images (AAO 2023)
Akshay Reddy1 BS, Nathaniel Tak2 BS, Jason I Dagoon1 BS, Parsa Riazi Esfahani1 BS, Muhammad Ghauri1 BS, Neel Nawathey3 BS, James B Martel4 M.D
1.Department of Medicine, California University Science of Medicine School of Medicine 2.Department of Medicine, Midwestern University Arizona College of Osteopathic Medicine 3.Department of Medicine, Touro College of Osteopathic Medicine 4.Department of Ophthalmology, California Northstate University College of Medicine
This study improves glaucoma diagnosis with machine learning. Retinal images from Kaggle dataset (1007 glaucoma & 1074 normal) were used to develop a DNN model on Tensorflow. The model was evaluated with AUC. precision, and recall. and trained for 2h 37m, achieving high AUC. recall, and precision.