AI powered pathological myopia diagnosis system

Abstract

This final year project addresses the critical challenge of pathological myopia diagnosis in Uganda, particularly in Kampala, where approximately 3% of the population is affected by this progressive vision-threatening condition. Pathological myopia, characterized by excessive axial elongation of the eye, typically begins between ages 6-13 and continues to progress throughout life, often leading to significant vision impairment during individuals' most productive years. Our project develops and implements an artificial intelligence (AI) system specifically designed to enhance the accuracy and efficiency of pathological myopia diagnosis in resource-constrained clinical settings. The system utilizes deep learning algorithms trained on fundus images to detect characteristic retinal changes associated with pathological myopia, including posterior staphyloma, chorioretinal atrophy, and myopic maculopathy. Through systematic development phases including requirements analysis, system design, model training, and validation testing, we created a user-friendly diagnostic support tool that achieves 89% accuracy in detecting pathological myopia markers. The system functions as a clinical decision support tool, providing ophthalmologists with AI-assisted analysis while maintaining medical professional oversight in the diagnostic process. Field testing conducted at Makerere University Hospital demonstrated the system's effectiveness in reducing diagnostic time by 64% compared to conventional methods, while maintaining high diagnostic accuracy. Healthcare providers reported high satisfaction with the tool's usability and integration into existing clinical workflows. This project contributes a practical, cost-effective solution to improve pathological myopia diagnosis in Uganda, with potential for broader application across similar healthcare settings in sub-Saharan Africa. The implementation of this AI diagnostic system represents a significant step toward modernizing eye care services and reducing preventable vision loss in affected populations