Smart alert: drowsiness detection system for enhanced safety

Drowsiness Detection System is developed to deliver a road safety product that is capable of recognizing behavioral states of driver fatigue using computer vision and machine learning techniques. As drowsiness contributes to a large portion of traffic accidents, it is important to monitor a driver's alertness in real time. The system uses a normal camera to record live video and observes movements of the eyes and head to recognize drowsiness. OpenCV will handle facial landmark detection and real-time video processing, while Dlib provides improved predictions on eye state classification. The Eye Aspect Ratio (EAR) method calculates how long a driver's eyes have been closed to distinguish between normal blink and prolonged eye closure, which may indicate drowsiness. In case of drowsiness, a warning alarm is triggered using Pygame, while the EAR falls below the defined threshold for a certain time.The development of the test has yielded promising results with reasonable accuracy under standard conditions with minimal violations in response time, although it may suffer from poor performances with bad lighting conditions and partial facial occlusions.