Drowsiness Detection System Using Image Analysis

Drowsiness while driving continues to be one of the main causes of severe road accidents and fatalities around the world. While traditional drowsiness detection systems that rely on physiological sensors are highly accurate, they tend to be invasive and not practical for regular use in vehicles. In this paper, we introduce a non-intrusive, image-based detection system that uses facial analysis to assess a driver’s level of alertness. By analyzing facial landmarks, the system calculates the Eye Aspect Ratio (EAR) and Mouth Aspect Ratio (MAR), which are strong indicators of eye closure and yawning, respectively. OpenCV and Dlib libraries are used to detect facial landmarks in real-time, and a Convolutional Neural Network (CNN) model is trained using annotated datasets from Kaggle and Roboflow. Our experiments show that while static image analysis has limitations due to changing environmental conditions, it holds significant promise for enhancing road safety through real-time monitoring. Future improvements will focus on making the system more robust and adaptable in real-world settings by incorporating dynamic tracking and IoT integration.