Gaze Direction Estimation Based on Natural Head Movements

In this paper, we present a new approach for estimating head pose and gaze direction under a monocular vision. The human visual line-of-sight consists of two components: the pose of head and the orientation of eyes within their sockets. First we compute the head pose of a user, and then calculate the gaze direction from the relative positions of the iris and the inner eye corner. The experimental results show that under natural head movements (yaw and pitch angles of head in real image sequences are up to 30deg and 20deg, respectively), our method achieves an average deviation to 2deg and 4deg in azimuth and elevation angles respectively.

[1]  Dmitry B. Goldgof,et al.  Matching point features under small nonrigid motion , 2001, Pattern Recognit..

[2]  Jan Flusser,et al.  Image registration methods: a survey , 2003, Image Vis. Comput..

[3]  Fang-Hsuan Cheng,et al.  Fast algorithm for point pattern matching: Invariant to translations, rotations and scale changes , 1997, Pattern Recognit..

[4]  Jian-Gang Wang,et al.  Eye gaze estimation from a single image of one eye , 2003, Proceedings Ninth IEEE International Conference on Computer Vision.

[5]  Alexander Zelinsky,et al.  An algorithm for real-time stereo vision implementation of head pose and gaze direction measurement , 2000, Proceedings Fourth IEEE International Conference on Automatic Face and Gesture Recognition (Cat. No. PR00580).

[6]  Christopher G. Harris,et al.  A Combined Corner and Edge Detector , 1988, Alvey Vision Conference.

[7]  Fang-Hsuan Cheng,et al.  Point pattern matching algorithm invariant to geometrical transformation and distortion , 1996, Pattern Recognit. Lett..

[8]  B. N. Chatterji,et al.  An FFT-based technique for translation, rotation, and scale-invariant image registration , 1996, IEEE Trans. Image Process..

[9]  Jan Flusser,et al.  A moment-based approach to registration of images with affine geometric distortion , 1994, IEEE Trans. Geosci. Remote. Sens..

[10]  Stephen E. Reichenbach,et al.  Two-dimensional cubic convolution , 2003, IEEE Trans. Image Process..

[11]  Yutaka Satoh,et al.  Using selective correlation coefficient for robust image registration , 2003, Pattern Recognit..

[12]  Jie Zhu,et al.  Subpixel eye gaze tracking , 2002, Proceedings of Fifth IEEE International Conference on Automatic Face Gesture Recognition.

[13]  Li Zhang,et al.  Robust face alignment based on local texture classifiers , 2005, IEEE International Conference on Image Processing 2005.

[14]  R. Keys.  Cubic convolution interpolation for digital image processing , 1981 .

[15]  Shinjiro Kawato,et al.  Gaze Direction Estimation with a Single Camera Based on Four Reference Points and Three Calibration Images , 2006, ACCV.

[16]  Takeo Kanade,et al.  Shape and motion from image streams under orthography: a factorization method , 1992, International Journal of Computer Vision.