Computer Vision for Automated Facial Characteristics Detection

Diagnosing diseases in their early stages allows people to see a specialist doctor before the disease reaches advanced stages and avoid any future complications. The use of a real-time imaging system to reliably give various information about the patient's condition, including gender classification, pterygium, and Bell's paralysis, contributes to reducing the duration of diagnosis and human errors. This study focuses on the use of three artificial intelligence algorithms based on deep learning, namely Visual Geometry Group (VGG16), Vision Transformer (ViT), and Xception, and evaluates their performance in detecting gender, pterygium, and Bell's paralysis. VIT has the highest overall performance results from the rest of the algorithms.