Abstract:
"Glaucoma is the second leading cause of blindness globally, according to the World Health Organization (WHO), making early detection crucial for preventing irreversible vision loss. Traditional diagnostic methods, while widely used, often suffer from limited accuracy and efficiency, especially in detecting early stages. This research proposes an innovative solution to this challenge by developing a machine learning-based system for glaucoma stage detection. By combining Convolutional Neural Networks (CNNs) with XGBoost and enhancing it with Explainable AI (XAI) techniques, this system offers improved precision and interpretability in diagnosing glaucoma.
The fusion of CNN's image processing capabilities with XGBoost's classification strength allows the model to accurately detect and classify glaucoma stages. Explainable AI further empowers healthcare professionals by providing interpretable insights into the model's predictions, building trust in the system and facilitating informed clinical decision-making. This novel approach holds significant potential for practical application in healthcare, as it enhances both diagnostic accuracy and transparency.
Test results reveal the system's high performance, achieving an accuracy rate of 81%, with strong precision, recall, and f1-score metrics across different glaucoma stages. These findings underscore the model's robustness and potential to outperform traditional diagnostic tools. The research addresses a critical gap in glaucoma detection, offering a new standard for applying deep learning in ophthalmology. Ultimately, this solution aims to revolutionize patient care, providing a valuable tool for early diagnosis and improved management of glaucoma, potentially benefiting millions of people worldwide."
