A Machine Learning-Based Framework for Early Detection of Trachoma in Resource-Limited Settings

Trachoma, a leading cause of infectious blindness, predominantly affects populations in resource-limited regionsworldwide. Early detection and intervention are critical to preventing disease progression and subsequent blindness.This study explores the application of machine learning techniques, including Random Forest, Support VectorMachine (SVM), Convolutional Neural Networks (CNNs), and K-Nearest Neighbors (KNN), for the identification ofTrachoma from image datasets. The research methodology encompasses data collection, preprocessing, featureextraction using CNNs, and the training and evaluation of various models. Among these, the SVM model achievedthe best performance with 70% accuracy, a precision of 0.68, a recall of 0.69, an F1-score of 0.68, and an ROC AUCof 0.72. These findings highlight the SVM model’s effectiveness and its promise as a tool for assisting healthcareprofessionals in making accurate and timely diagnoses of Trachoma