Abstract—According to the previous researches by experts, the smoking behavior in driving will cause three hazards: reducing vision, distracting, and irritating. Thus the risk is extremely high. By detecting whether there is any object of a cigarette, the smoking behavior in driving is recognized, and it will greatly help for driving safety. In this paper, the YOLOv2 deep-learning image based methodology is applied for driver’s cigarette object detection. The driver's images are captured by a dual-mode visible light and near-infrared camera, and the developed system judges whether or not there is driver smoking behavior in the day and night conditions. By the YOLOv2 deep learning network, the pre-prepared images of driver smoking behavior are marked, and the cigarette detector is trained to detect the cigarette object when the driver is smoking. In experimental results, the applied deep learning based design performs that the precision is up to 97% and the recall is 98%. Besides, by the proposed design, the average accuracy of cigarette detection is up to 96% during the day condition, and that of cigarette detection is up to 85% during the night condition.
Index Terms—smoking behavior detection, deep learning, YOLOv2, driving safety
Cite: Tzu-Chih Chien, Chieh-Chuan Lin, and Chih-Peng Fan, "Deep Learning Based Driver Smoking Behavior Detection for Driving Safety," Journal of Image and Graphics, Vol. 8, No. 1, pp. 15-20, March 2020. doi: 10.18178/joig.8.1.15-20
Copyright © 2020 by the authors. This is an open access article distributed under the Creative Commons Attribution License (CC BY-NC-ND 4.0), which permits use, distribution and reproduction in any medium, provided that the article is properly cited, the use is non-commercial and no modifications or adaptations are made.
Copyright © 2012-2020 Journal of Image and Graphics, All Rights Reserved