Guarding Vigilance: A Smart System to Detect and Defeat Drowsiness in Real-Time

Driver drowsiness is a leading cause of road accidents, often resulting in severe consequences. This research paper presents a real-time, non-intrusive driver drowsiness detection system that leverages the Eye Aspect Ratio (EAR) derived from eye movement tracking to evaluate fatigue levels. By analyzing facial landmarks using advanced computer vision techniques via a webcam or vehicle-mounted camera, the system detects when the EAR drops below a customizable threshold and triggers an alert to warn the driver. The non-invasive design ensures comfort and ease of integration into vehicles, while the model's training on a large dataset ensures accurate detection across diverse drivers. Real-time testing demonstrates the system's effectiveness in improving road safety, and future enhancements, including adaptive thresholding, multi-modal data fusion, and real-time calibration, are proposed to further optimize performance and reliability under varying conditions. Key Words: computer vision, drowsiness detection, EAR, real-time testing.