Gaze Detection Based on Head Pose Estimation in Smart TV

Abstract

In this paper, we propose a new gaze detection method based on head pose estimation in smart TV. First, the face, eyes and center of both nostrils positions are detected by using Adaboost method. After that, the proposed method uses the template matching method to find the position of the user's shoulder as well as the left and right boundaries of the face. Based on the positions of the detected shoulder and face's boundary, the horizontal and vertical head poses are estimated. Experimental results show that our method shows a good performance in user's gaze detection.