Gaze Tracking

This paper proposes a depth compensation model to reduce parallax error for head-mounted eye trackers. The model assumes the distance between the user and the target is known and adjusts the gaze estimation accordingly. The model is based on geometric principles and is tested in a controlled environment. The paper also presents a binocular gaze estimation method using epipolar geometry. The results show that the depth compensation model can significantly reduce the parallax error in different depth planes.