Research on Eye-state Based Monitoring for Drivers´ Dozing

In the paper, an eye-state based doze monitoring method is proposed, which sets improvements in face detection and doze monitoring. Although having been researched intensively in face detection, multi-pose problem still remains to be solved, so a new multi-pose oriented AdaBoost algorithm for face detection is set forth to enhance the detection accuracy. According to the practical application, the paper also proposes a new eye-state recognition algorithm, in which two eye features are quantified and fused to make a more accurate detecting result. The experiments show that the method has its excellence both in real-time and accuracy performance for drivers´ dozing.