3D Gaze Estimation Method Based on Binocular Model
-
摘要: 提出一种基于双眼模型的三维视线估计方法,以期克服二维视线跟踪方法固有的两个缺点,即对于使用者头部运动限制以及个体参数标定过程繁琐。该方法首先利用两个安装有红外灯组的摄像机组成立体视觉系统,实现人脸双眼图像的采集,之后,基于眼睛角膜球状假设,建立了一个适用于三维视线估计的三维眼模型,在此基础上,推导出人眼三维光轴的估计方法。用左右眼光轴与屏幕交点连线的中点估计方法,计算出注视点。该方法可在使用者头部运动的情况下,完成视线估计,且只需进行一次系统校准即可使用,无需其它标定过程,视线估计精度较高。Abstract: This paper presents the 3D gaze estimation method based on binocular model. The method can overcome the two inherent problems of the 2D gaze estimation method, namely restriction on head movement and complex procedures for calibrating individual parameters. Firstly, we use two cameras on which groups of infrared lights are installed around the center of the lens to establish a stereo vision system. Then using this stereo vision system, the cameras capture eye images. Secondly, based on the non-spherical model of a cornea, we set up the 3D eye model which is adaptive to estimate 3D gaze. Thirdly the binocular optical axis calculation method is deduced. Finally, the midpoint between two intersection points of the left and right eye optical axis and the computer screen is calculated. With these, the fixation point is determined. The 3D gaze estimation method can estimate the eye gaze under natural head movement. Moreover, except only once, calibration by a user is not required (no calibration). The 3D gaze estimation method has high accuracy.
-
Key words:
- binocular vision /
- calculations /
- calibration
-
[1] Duchowski A T. Eye tracking methodology: theory and practice[M]. London:Springer-Verlag,2003 [2] Shao G J, Che M, Zhang B Y, et al. A novel simple 2D model of eye gaze estimation[C]//Intelligent Human-Machine Systems and Cybernetics,2010:300-304 [3] 周园,赵歆波,张昀.针对真实场景的实时视线跟踪系统设计[J].机械科学与技术,2010,29(2) Zhou Y, Zhao X B, Zhang Y. Design of real-time eye-gaze tracking system for actual scenario[J]. Mechanical Science and Technology for Aerospace Engineering,2010,29(2) (in Chinese) [4] Shih S W, Liu J. A novel approach to 3-D gaze tracking using stereo cameras[C]//IEEE Transactions on Systems, Man, and Cybernetics,2004,B:234-245 [5] Shih S W, Wu Y T, Liu J. A calibration-free gaze tracking technique[C]//Proceedings of International Conference on Pattern Recognition,2000,4:201-204 [6] Kohlbecher S, Bardinst S, Bartl K, et al. Calibrationfree eye tracking by reconstruction of the pupil ellipse in 3D space[C]// Proceedings of the 2008 Symposium on Eye Tracking Research & Applications,2008:135-138 [7] Yamazoe H, Utsumi A, Yonezawa T, et al. Remote gaze estimation with a single camera based on facial feature tracking without special calibration actions[C]//Proceedings of the 2008 Symposium on Eye Tracking Research & Applications,2008:245-250 [8] Sugano Y, Matsushita Y, Sato Y. Calibration-free gaze sensing using saliency maps[C]//IEEE Conference on Computer Vision and Pattern Recognition,2010,2667-2674 [9] Noris B, Keller J B, Billard A G. Gaze direction detection for lay users in unconstrained environments[J]. Computer Vision and Image Understanding,2011,115:476-486 [10] Noris B K, Benmachiche A G. Billard calibration-free eye gaze direction detection with gaussian processes[C]//Proceedings of the International Conference on Computer Vision Theory and Applications,2008 [11] Gielen C, Dijkstra T, Roozen I J, et al. Coordination of gaze and hand movements for tracking and tracing in 3D[J]. Cortex,2009,45:340-355 [12] Chen J, Tong Y, Gray W, et al. A robust 3D eye gaze tracking system using noise reduction[C]//Proc. of Eye Tracking Research & Applications Symp,2008:189-196 [13] Pogalin E, Redertl A, Patras I, et al. Gaze tracking by using factorized likelihoods particle filtering and stereo vision[C]//3D Data Processing, Visualization and Transmission,2006:57-64 [14] Osaka R. Experimental psychology of eye movements (in Japanese)[M]. The University of Nagoya Press, 1993
点击查看大图
计量
- 文章访问数: 148
- HTML全文浏览量: 37
- PDF下载量: 7
- 被引次数: 0