Performance Analysis of Deep Learning Models for Accurate Glaucoma Detection

The realm of healthcare, machine learning has emerged as an especially profound, is easy to use and fairly accurate method for easy solutions and making medical procedures more and more accurate. This research study focuses on glaucoma, which is one of the more common and yet less spoken of visual disorders in the country and in the world. Seeing the predicted increased load on the health system with this disease we propose of flurry of comparison between potential machine learning solutions and deep learning models that can detect easily and accurate glaucoma within patients reducing the doctor's workload is proposed. The potential models picked for this work are inceptionv3, ResNet50, VGG16, VGG19, Xception and a custom convolutional neural network (CCNN) as well. The obtained accuracy (in %) for glaucoma detection is 99.43 using VGG19. This may provide an effective application and solution for glaucoma detection.