Drivers Fatigue Detection Using EfficientDet In Advanced Driver Assistance Systems

Modern cars have focused on road safety by guaranteeing driver, pedestrians, and other traffic object safety. Advanced driver assistance systems are a set of intelligent systems that support the driver by providing more information about the surrounding environment. The fatigue detection system is an intelligent system that detects the driver's face and gauges the driver's tiredness state. Such a system can prevent accidents by stropping the car if the driver is drowsy. In this paper, we propose a driver fatigue detection based on object detection model fatigue indicators. The efficientDet model was used to detect the state of the eye and mouth states then the eyes' closure duration/Percentage of eye closure (PERCLOS) and yawning frequency/frequency of mouth (FOM) were used to judge fatigue state. The efficientDet is a lightweight object detection model with high performance. The proposed approach was evaluated on the National Tsing Hua University Driver Drowsiness Detection (NTHU-DDD) dataset. The proposed approach has achieved an accuracy of 96.05% and real-time processing. The reported results show the efficiency of the proposed model for driver fatigue detection.