On Driver Eye Closure Recognition for Commercial Vehicles

This paper addresses the issue of driving while drowsy and proposes a passive eye monitoring-based driver eye closure recognition system. It reviews the core algorithmic building blocks of this system along with in-depth analysis of operational test field characteristics. The system operates equally in both day and night-time. It automatically finds the drivers eyes in the images, tracks the eye location in a wide range of head and eye motion, and estimates in real-time the eye state as either open or closed eye, and further infers to driver drowsiness state. This paper reports as well the experimental results on a diverse and challenging set of subjects and environmental driving conditions.