AI-Based Driver Drowsiness Detection System Using Computer Vision for Road Safety

Abstract

Driver drowsiness is a major contributor to road traffic accidents worldwide, posing significant risks to public safety and resulting in substantial economic and social losses. Traditional approaches to detecting driver fatigue often rely on physiological sensors or vehicle-based indicators, which may require specialized hardware, intrusive installation, or exhibit reduced reliability under varying driving conditions. This research presents an AI-based driver drowsiness detection system using computer vision to provide a non-intrusive, real-time solution for enhancing road safety. The proposed system continuously monitors the driver's facial features through a camera and applies computer vision techniques to detect signs of fatigue, including prolonged eye closure, frequent blinking, yawning, and abnormal head movements. Facial landmarks are extracted using MediaPipe Face Mesh, enabling the computation of key fatigue indicators such as Eye Aspect Ratio (EAR) and Mouth Aspect Ratio (MAR). These features are analyzed by an artificial intelligence model to classify the driver's alertness level into normal, warning, or drowsy states. When drowsiness is detected beyond predefined thresholds, the system immediately activates an audio-visual alert to regain the driver's attention and reduce the likelihood of accidents. The proposed framework is designed for real-time operation with low computational complexity, making it suitable for deployment on embedded platforms and integration into advanced driver-assistance systems (ADAS). Experimental evaluation demonstrates that the proposed approach achieves high detection accuracy while maintaining reliable performance under diverse lighting conditions and varying facial expressions. The combination of facial landmark analysis, computer vision, and artificial intelligence offers a cost-effective, scalable, and practical solution for intelligent transportation systems. The proposed framework contributes to the advancement of AI-enabled road safety technologies by providing an efficient, non-contact, and real-time driver monitoring system capable of reducing fatigue-related accidents and supporting the development of safer smart mobility infrastructures.