Early Detection and Classification of Cataracts Using Smartphone Imagery Based on Support Vector Machine (SVM) and Certainly Factor Methods

Cataracts are one of the leading causes of blindness worldwide. Current technological advancements, especially in Artificial Intelligence (AI), offer the potential to develop products that facilitate the detection of cataracts. The implementation of AI in portable handheld devices is a significant technological development. Therefore, this research aims to create a system for early detection and classification of cataract disease using AI with smartphone images. The focus of the research will be on developing better detection and classification methods, which will work in tandem with the preprocessing of input images. The approach involves using Machine Learning to detect physical symptoms and an expert system based on the certainty factor to identify patient symptoms. The detection of eye cloudiness is divided into three stages: image preprocessing using pixel refilling, pupil segmentation using the Circle Hough Transform, and cataract detection using Support Vector Machine (SVM). The use of the SVM method allows for faster and more accurate detection with a smaller dataset required to build the system. The expert system method will be developed with the assistance of ophthalmologists, aiming to mimic the diagnostic process of experts through confidence values. The accuracy of the system using cross-validation methods is 90.15% with a standard deviation of 5.48%.