Detection of Driver Gaze Direction and Considerations on its Application to Glance Away Detection using Image Processing

摘要

To realize driving support systems friendly to drivers, it is important to detect driver states (dozing, glancing away, etc.). This paper describes the development of a technique to detect driver gaze direction during driving by time-varying image processing. In this method to detect the gaze direction, the facial direction and the eyeball direction can be detected separately. Verification experiments proved that the developed method is applicable to glance away detection. Moreover, verification experiments concerning glance away behavior were performed using a driving simulator. Some information effective for system development was obtained.