Driver Somnolence and Lethargy Detection Model based on Eyes and Facial Movement using OpenCV Library (EyeF Model)

Fatigue and drowsiness are significant contributing factors to driver alertness, which tend to cause severe road accidents because reaction times are slower and judgment is impaired. It is important to detect these early to ensure safe driving. Monitoring eye movements has been found to be a good method of identifying drowsiness signs, with methods such as PERCLOS (Percentage of Eyelid Closure), EAR (Eye Aspect Ratio), and YOLO (You Only Look Once) being used to monitor in real-time. Yet, each of these approaches possesses its own difficulties, including compensation for lighting variation and separating typical behaviour from tiredness. Through the integration of the strengths of all three approaches, a hybrid model assists in eliminating these drawbacks and providing a more balanced and more precise method of identifying indicators of fatigue. Testing shows that this method delivers high precision in real-time, significantly reduces false alerts, and improves overall system reliability. With its strong potential for use in everyday transportation, this approach could make driver monitoring more effective and help prevent accidents caused by fatigue.