|
|
 |
| |
|
|
|
Head Pose and Gaze Direction Tracking for Detecting a Drowsy Driver |
|
|
|
PP: 505-512 |
|
|
|
Author(s) |
|
|
|
In-Ho Choi,
Yong-Guk Kim,
|
|
|
|
Abstract |
|
|
| This paper proposes a new drowsy-driver detection system that uses the head-pose, gaze direction and eye-blinking states
of a driving person. Head-pose of the driver is estimated by using optical flow of the facial features, which are acquired with a corner
detection algorithm. To estimate the gaze direction of the driver, we trace the center point of the pupil. The eye-blinking is estimated by
using the Integral Projection Function (IPF). Then the level of drowsiness is calculated by combining these three parameters in a data
fusion method. Performance of the system is evaluated using the new driver database. Result suggests that the system has a practical
potential for detecting drowsiness of the drivers. |
|
|
|
|
|
|
|
