Drowsiness Detection System for Safe Driving

Driver drowsiness is a major contributor to road accidents, demanding continuous monitoring systems for enhanced road safety. This paper presents a real-time, camera-based drowsiness detection system using computer vision and machine learning techniques. The model extracts facial landmarks to compute Eye Aspect Ratio (EAR) for eye closure detection and Mouth Aspect Ratio (MAR) for yawn identification. EAR and MAR thresholds are analyzed with head pose estimation to determine the severity of driver fatigue. A Convolution Neural Network (CNN) further improves classification accuracy under varying illumination and driving conditions. When abnormal EAR and MAR values are detected, the system issues an instant alert to prevent accidents. Experimental results show high accuracy and low latency, making the proposed method suitable for integration into intelligent transportation systems.