Human Robot Interaction using Face Pose Recognition

Face pose-based robot control plays an important role in natural human-robot interaction. In this paper, we propose a robust face pose estimation method based on manifold learning to control the wheeled robot. We represent each pose of a person's face as a connected low-dimensional appearance manifolds which are approximated by affine plane. Pose recognition task is to find sub-pose by computing the minimal distance from the given face image to sub-pose manifold. Based on the pose recognition result, the robot is controlled. Initial experiments are promising.