Glaucoma Identification Using AI-based Deep Learning Models

Early detection is key to effectively managing glaucoma, the leading cause of irreversible blindness. This study introduces a deep learning model based on Convolutional Neural Networks (CNNs) designed to automatically detect glaucoma using retinal images. The model's architecture includes several layers that help it learn and identify signs of glaucoma. To make the model more reliable and widely applicable, a lot of data was processed and enhanced. Key measures used to evaluate the model's performance included the AUC-ROC curve, F1-Score, accuracy, sensitivity, specificity, and precision. When tested on a separate dataset, the model outperformed traditional eye care methods and existing models, achieving an AUC-ROC of 0.978, sensitivity of 93.8%, specificity of 96.5%, precision of 94.1%, F1-Score of 94.0%, and accuracy of 95.2%. The proposed deep learning model offers a reliable, accurate, and scalable solution for glaucoma screening, with the potential to improve early detection and patient outcomes in clinical settings.