Automatic Detection of Driver Fatigue - Phase 1

Project summary:

We plan to build a system that tracks the eyes of the driver (i.e., the iris) in order to detect fatigue. In particular, fatigue can be detected by visually monitoring the eyes and the eye blink rate on the head of a driver. It should be stated that iris pattern recognition- (e.g., IriScan system) has become very popular as an effective biometric means of identification the recent years. Standard monochrome cameras can detect an iris at distances of 30-45 cm. Since the shape of the iris does not change significantly from person to person, deformable models can be used to detect and track the iris. In addition, the eyes move according to several patterns that can be captured by the deformable models. Issues such as the false alarms rate should be addressed in order to build a reliable fatigue detection system.