1. Detecting Distracted Driving with Deep Learning
- Author
-
Ofonime Dominic Okon and Li Meng
- Subjects
business.industry ,Accident prevention ,Computer science ,Deep learning ,Speech recognition ,02 engineering and technology ,01 natural sciences ,Convolutional neural network ,Task (project management) ,010309 optics ,020210 optoelectronics & photonics ,Triplet loss ,Distraction ,0103 physical sciences ,Pattern recognition (psychology) ,0202 electrical engineering, electronic engineering, information engineering ,Distracted driving ,Artificial intelligence ,business - Abstract
Driver distraction is the leading factor in most car crashes and near-crashes. This paper discusses the types, causes and impacts of distracted driving. A deep learning approach is then presented for the detection of such driving behaviors using images of the driver, where an enhancement has been made to a standard convolutional neural network (CNN). Experimental results on Kaggle challenge dataset have confirmed the capability of a convolutional neural network (CNN) in this complicated computer vision task and illustrated the contribution of the CNN enhancement to a better pattern recognition accuracy.
- Published
- 2017
- Full Text
- View/download PDF