Researchers develop AI system that can identify when people are texting while driving

Researchers from the University of Waterloo have developed a software that can identify actions by drivers that can lead to an accident. Cameras and artificial intelligence are used in tandem to identify actions such as texting, talking on the phone, or reaching into the back of the vehicle to retrieve something. Once the action is identified, it is also graded and classified according to the severity of the safety threat.[caption id="attachment_4032869" align="alignleft" width="380"] Image: AP[/caption]The system can be used to warn the driver of dangerous distractions, or even take over control of the vehicle if needed. Fakhri Karray, a professor at the University of Waterloo says, "The car could actually take over driving if there was imminent danger, even for a short while, in order to avoid crashes." The system can be used to improve road safety, and alert the drivers when they are dangerously distracted.The system was an evolution of the work being done at the Centre for Pattern Analysis and Machine Intelligence (CPAMI) in Waterloo. Another system being developed at the same facility tracks eye movements to identify when the driver is liable to fall asleep, and if they are paying sufficient attention to the road. The findings of the study were presented at the International Conference on Image Analysis and Recognition in Montreal, Canada.