2010

Contact: Bridgette L. Gollinger
DENSO International America, Inc.
Phone: (248) 372-8266
bridgette_gollinger@denso-diam.com

SOUTHFIELD, Mich. <#843;> The U.S. National Highway Traffic Safety Administration (NHTSA) estimates that about 100,000 of reported crashes occur as a result of driver drowsiness or fatigue. What's even more alarming is that NHTSA considers sleep-deprived drivers a hazard equal in severity to drunk drivers. The greater the sleep deprivation, the closer the correlation to higher levels of intoxication.

Although vehicle active safety systems have helped decrease the number of deaths occurring in traffic accidents, the number of traffic accidents is still increasing. In response to this, DENSO has developed a method to detect driver drowsiness using the whole facial image, including information related to the eyes. The drowsiness detection system uses a camera located on the dashboard, an image processing algorithm, and a drowsiness detection algorithm.

Brain wave detection, which is commonly used to detect drowsiness, is not suitable for the in-vehicle system because it uses sensors that must be attached to the head. DENSO's method does not require the attachment of any contact-type sensors. Since the facial nerve nucleus is contained in the brain stem, which is defined as an organ of drowsiness, DENSO has adopted facial expression as the drowsiness index. Also, DENSO examinations indicated facial expressions have the highest correlation with brain waves and it detects drowsiness at an earlier stage than other indices considered suitable for drowsiness detection, such as the percentage of eye closure time, heart rate, and lane deviation.

DENSO's method categorizes drowsiness into 6 levels: 1)not sleepy 2)slightly sleepy 3)sleepy 4)rather sleepy 5)very sleepy, and 6)sleeping by using features of facial expression based on the mechanism of facial muscle activities. To measure the driver's drowsiness, DENSO developed a method that detects three-dimensional coordinates of 68 points on the driver's face frame from the driver's video-captured facial image.

DENSO is continuing to work to ensure the accuracy of this drowsiness detection method by verifying practical effectiveness using a motion-based driving simulator and/or real car examinations. DENSO has also started to develop a feedback system that keeps the driver awake effectively and naturally.

# # #