This paper describes a noncontact man-machine interface based on face orientation. Real-time images of an operator's face were observed by a USB camera and changes in the dark area of the nostrils were utilized for the recognition of face orientation. When the operator faced up, dark areas of both nostrils increased in area, and when the operator faced down, such dark areas decreased, respectively. In contrast, the difference between nostril areas could be caused when the face was turn to the side. Here, these characteristics were reflected in face-orientation recognition. The interface we developed was applied to electrical wheelchair operation.
展开▼
