Abstract
The traffic accident occurs due to aimless driving such as the driver's looking aside, fatigue, and a nap. In this research, we aim at development a system that can judge the aimless driving by driver's facial feature point. Usually RGB images are used for facial feature point's detection. But they are influenced by environment light, so in the car with a large change of environment light, the detection of the stable facial feature point is difficult. In this paper, we use depth information that is not influenced by changing light for facial and nose region detection.
