Precise eye location in driver fatigue state surveillance system

Whether a driver is fatigued or not can be reflected by his eyes during the course of driving, so it is practical to use the information about the driver´s eyes to monitor his fatigue state. Among these, eye location is the key problem. Based on the existing eye location algorithms, specifically for the driver fatigue state surveillance system, this paper presents a method based on facial features. The experiment results show that this method is robust and real-time, it´s adaptive to human face in different conditions such as different backgrounds, different light conditions, different size and rotating angles, open and close eyes, with glasses on or not.