Anzeige
Erschienen in:
2017 | OriginalPaper | Buchkapitel
Estimating 3D Gaze Point on Object Using Stereo Scene Cameras
verfasst von : Zhonghua Wan, Caihua Xiong
Erschienen in: Intelligent Robotics and Applications
Aktivieren Sie unsere intelligente Suche, um passende Fachinhalte oder Patente zu finden.
Wählen Sie Textabschnitte aus um mit Künstlicher Intelligenz passenden Patente zu finden. powered by
Markieren Sie Textabschnitte, um KI-gestützt weitere passende Inhalte zu finden. powered by
Abstract
3D eye gaze estimation in real environment is still challenging. A novel method of scene-based 3D gaze estimation is proposed in this paper. As this model combines two models, the 2D nonlinear polynomial mapping model of traditional regression-based gaze estimation and the 3D visual axis linear ray model of traditional geometry-based gaze estimation, it includes two steps. The first step is to estimate the visual axis from the pupil center in an eye camera image. The second one is to estimate the 3D gaze point which is the intersection between the visual axis and the scene object, which can be obtained by stereo scene cameras. As the 3D gaze points are on the object, rather than outside or inside the object like geometry-based 3D gaze estimation, this method is potential for human robot interaction in real environment. Through a simple test, the accuracy of our 3D gaze estimation system is acceptable.