A Real-time Driver Fatigue Detection Method Based on Two-Stage Convolutional Neural Network. Issue 2 (2020)

• Record Type: Journal Article
• Authors: He, Hu; Zhang, Xiaoyong; Jiang, Fu; Wang, Chenglong; Yang, Yingze; Liu, Weirong; Peng, Jun
• Is Part Of: IFAC-PapersOnLine. Volume 53:Issue 2(2020)
• Page Start: 15374
• Page End: 15379
• Publication Date: 2020
• Subjects: Driving safety -- Driver fatigue detection -- Facial feature -- Convolutional neural network -- Location Detection Network -- State Recognition Network; Automatic control -- Periodicals; 629.805
• DOI: 10.1016/j.ifacol.2020.12.2357
• Languages: English
• ISSNs: 2405-8963

Abstract

Abstract: Fatigue-related traffic accidents have a higher mortality rate and cause more significant damage to the environment. To ensure driving safety, a real-time driver fatigue detection method based on convolutional neural network (CNN) is proposed in this paper. The proposed fatigue driving detection method is cascaded by two CNN-based stages, including a detecting phase and classifying phase. The Location Detection Network is designed to extract facial features and localize the driver's eyes and mouth regions. Then the State Recognition Network is training to recognize the driver's eyes and mouth status. Simulations show that the proposed method has good effect of real time process and high accuracy of detection. Experiments conducted on Raspberry Pi 4 embedded system indicate that the proposed method has a good performance in the real driving environment.