Data Pipeline Design for Dangerous Driving Behavior Detection System.

More than half of all traffic accidents are caused by non-compliance with safe driving duties. If the driver is awarded for carefully following safety regulations, the number of non-compliance with safety practices will drop significantly. Additionally, research is being conducted to reduce the number of cases by identifying driving behavior using data generated from a front camera of a vehicle. However, as data sources increase, there are difficulties in data analysis for artificial intelligence models and data science research. In this paper, we propose a data pipeline design for a dangerous driving behavior detection system. The proposed system is an data pipeline design that can assist refinement, processing, and analysis after collection of vehicle front camera data for learning dangerous driving behavior detection system.