Driver Distraction and Drowsiness Detection Based on Object Detection Using Deep Learning Algorithm

Abstract

Distracted driving is a major global contributing factor to traffic accidents. Distracted drivers are three times more likely to be involved in an accident than non-distracted drivers. This is why detecting driver distraction is essential to improving road safety. Several prior studies have proposed a range of methods for identifying driver distraction, including as image, sensor, and machine learning-based approaches. However, these methods have limitations in terms of accuracy, complexity, and real-time performance. By combining a convolutional neural network (CNN) with the You Only Look Once (YOLO) object identification method, this study suggests a unique way to driver distraction detection The two primary phases of the suggested paradigm are object identification utilizing Yolo and classification of the identified things. The YOLO algorithm is used to identify and pinpoint the driver's hands, face, and any other objects that might draw their attention away from the road. The objects that have been observed are then categorized using a CNN to determine whether or not the driver is distracted. When evaluated on a publicly available dataset, the proposed model shows good performance in detecting driver preoccupation. Utilize the CNN algorithm in addition to ocular features to determine the driver's level of fatigue. The proposed method might be incorporated into advanced driver assistance systems with real-time environment to improve road safety.