Big data and artificial intelligence (AI) to detect glaucoma

Glaucoma is an eye condition that is one of the most prevalent causes of blindness due to damage to the optic nerve. According to existing studies, it is the second most common reason for vision loss worldwide. With the advanced development of artificial intelligence (AI), it is necessary to develop a glaucoma detection system for diagnosis. Therefore, the primary objective of this paper is to create a system for detecting glaucoma using retinal fundus images, which can help determine if the patient was affected by glaucoma. Although several methods have been applied to detect glaucoma in the past decades, it is essential to use an advanced AI technique with a glaucoma detection system. Thus, in this paper, we divided our task into threefold: 1) segmentation, 2) classification, and 3) deployment. The U-Net architecture is implemented for segmentation. The pretrained GC-Net model is proposed for classification. Finally, based on the segmentation and classification, we developed a glaucoma detection system for diagnosis. This study uses ACRIMA datasets for training and testing. The result of this model is evaluated using various parameters such as accuracy, sensitivity, specificity, f1-score, and auc score. The output is compared to deep learning models such as ResNet, CNN, Inception V3, and TB-Net. The proposed model achieved 96% accuracy in training and 93% accuracy in testing. Overall, the performance of the proposed model is better in all the analyses. Copyright © 2022 IEEE.