Fatigue Driving Detection Based on Facial Features

In recent years, traffic accidents caused by fatigue driving are common. After a long period of continuous driving, the driver will have a dysfunction of mental and physiological functions, and his judgment and execution will be affected. The traditional fatigue detection methods are mainly based on physiological characteristics and vehicle driving information. These two detection methods are complex, require high detection accuracy and expensive equipment, and may be invasive to the driver's body, affecting the normal operation of the driver, and cannot be put into large-scale use in actual production. In this paper, we combine the methods of deep learning in the field of computer vision, based on the YOLOv5 object detection network model, using foreign public dataset YawnDD and CEW closed-eye dataset, and using multi-feature detection to detect the driver's eye and mouth fatigue features. In addition, this paper introduces a CBAM attention mechanism based on the YOLOv5 network structure to improve network performance and make fatigue detection more accurate.