Comprehensive Machine Learning-based Approach for Intelligent Glaucoma Detection

Glaucoma, a leading cause of irreversible blindness, requires early detection and timely intervention for effective management. With the advancements in Machine Learning (ML), an intelligent glaucoma detection system has emerged, offering a comprehensive end-to-end approach to tackle this sight-threatening condition. The system begins with the crucial step of data collection, where a diverse dataset of glaucoma-related images is gathered from reliable sources. The collected data then undergoes a pre-processing stage, where image enhancement techniques are applied to improve noise reduction, contrast enhancement, identify ROI, and remove blood vessels. Next, feature extraction using the Cup-to-Disc Ratio (CDR) and Inferior Superior Nasal Temporal Ratio (ISNT) ratio, and training ML models Support Vector Machine (SVM) and Adaptive Boosting (AdaBoost). The performance of different combinations of features and models is evaluated using metrics such as accuracy, specificity, sensitivity, precision, and F1 score. The results identified ISNTR with AdaBoost as the best combination, which is identified based on the highest performance metrics achieved. Finally, the selected ML model is deployed in a user-friendly web application or website. Users can easily access the website, upload their eye images, and obtain prompt predictions for the presence of glaucoma. This end-to-end approach provides a seamless and efficient solution for early glaucoma detection, allowing for timely intervention and improved patient outcomes.