Enhancing Driver Safety: Real-Time Drowsiness Detection through Eye Aspect Ratio and CNN-Based Eye State Analysis

As technology rapidly advances, our reliance on automobiles for transportation has grown significantly. However, this increased use of vehicles has brought about a critical concern for road safety - drowsy driving. To address this issue, we intro- duce a novel Drowsiness Detection System (DDS) that harnesses the capabilities of machine learning, specifically OpenCV and Keras models. Our DDS adopts a comprehensive approach to identify driver drowsiness by integrating computer vision techniques and deep learning algorithms. OpenCV is employed for real-time video analysis, tracking crucial facial and eye-related indicators such as eye closure, blink rate, and head position. Moreover, the heart of our DDS relies on deep learning models implemented through Keras, trained on a diverse dataset of driver behavior examples. These Keras models can dynamically adapt to varying lighting conditions, driver appearances, and different vehicle environments, ensuring reliable performance. The DDS promptly alerts the driver upon detecting signs of drowsiness, using both audible alerts and visual prompts.