Driver Fatigue Detection System Based on DSP Platform.

Due to non-invasiveness, monitoring driver state in computer vision (CV) has become a major way to detect driver fatigue. In contrast to other researches, we brought the driver fatigue detection system on the basis of DSP platform, which can make contribution to application on the integrated system for vehicle. However, the conventional system cannot easily be transplanted into DSP due to its storage and computation capacity. Therefore, designing an algorithm that can detect the fatigue efficiently is goal in this study. As the most important part of system, the geometric relationship and shape information within near frontal face is employed in the eye detection part, which depicts the eyebrow, eye and nose. In experiment part, a self-made database is assembled to test the performance of system. As the results of experiment, the detection rate of eye is achieved at 92.71 % and driver fatigue state is obtained at 97.5 %.