Driver Drowsiness Prediction Based on Multiple Aspects Using Image Processing Techniques

– Road accidents caused by driver drowsiness have been a significant concern for decades. With advancements in automation, various drowsiness detection systems have been developed, yet they still require improvements in efficiency, accuracy, cost-effectiveness, and speed. This paper proposes an integrated approach for driver drowsiness detection based on multiple aspects, utilizing eye and mouth closure status (PERCLOS) along with a newly introduced Facial Aspect Ratio (FAR), similar to EAR and MAR. This approach helps detect drowsiness indicators such as closed eyes, yawning, nodding, or covering the mouth, which are common human responses to fatigue. The system further integrates texture-based gradient patterns to identify drivers’ faces from different angles, recognize sunglasses, and detect scenarios where hands cover the eyes or mouth. By leveraging image processing techniques and deep learning algorithms, the proposed system aims to enhance real-time driver monitoring, ensuring improved road safety and reducing the risks associated with drowsy driving.