Method and apparatus for eye gaze tracking

1. An eye gaze tracking method comprising the steps of:

detecting a position of an iris center and a position of a corresponding eye corner of a user;

determining an eye vector associated with a gaze direction of the user using the detected iris center position and the corresponding eye corner position;

determining an initial static gaze point of the user by mapping the eye vector with a gaze target; and

applying the initial static gaze point to a head pose estimation model that is arranged to model a head pose of the user so as to devise a final gaze point of the user;

wherein the eye vector g is defined by the position of the iris center p_iris and the position of the eye corner p_corner via a relationship of g=p_corner−p_iris; and

wherein the head pose estimation model is arranged to apply a sinusoidal head model to simulate a shape of the head of the user.

2. An eye gaze tracking method in accordance with claim 1 , wherein the step of detecting the position of the iris center and the position of the corresponding eye corner includes the steps of:

detecting and extracting an eye region from the image of the user; and

detecting the position of the iris center and the position of the corresponding eye corner from the extracted eye region.

3. An eye gaze tracking method in accordance with claim 2 , wherein the step of detecting the position of the iris center and the position of the corresponding eye corner from the extracted eye region comprises the step of:

locating the position of the iris center using an eye iris center detection approach which combines intensity energy and edge strength information of the eye region; and

locating the position of the corresponding inner eye corner using an eye corner detection approach implemented in a multi-scale eye corner detector based on Curvature Scale Space and template match rechecking method.

4. An eye gaze tracking method in accordance with claim 1 , wherein the head pose estimation model includes an adaptive weighted facial features embedded in a POSIT (AWPOSIT) algorithm.

5. An eye gaze tracking method in accordance with claim 1 , wherein the eye corner is an inner eye corner.

6. An eye gaze tracking method in accordance with claim 1 , wherein the head pose estimation model is arranged to apply weighting to one or more facial features of an image of the user for head pose estimation.

7. An eye gaze tracking method in accordance with claim 6 , wherein the weighting depends on a normal direction of the one or more facial features.

8. An eye gaze tracking method in accordance with claim 6 , wherein the head pose estimation model is arranged to apply a larger weighting to an eye of the image of the user.

9. An eye gaze tracking method in accordance with claim 1 , further comprising the steps of:

presenting a gaze target for view by the user; and

determining an eye vector of the user for viewing the gaze target so as to map the eye vector with the gaze target.

10. An eye gaze tracking method in accordance with claim 9 , further comprising the step of:

repeating the gaze target presentation and eye vector determination steps using one or more different gaze targets.

11. An eye gaze tracking method in accordance with claim 1 , wherein the step of applying the initial static gaze point to a head pose estimation model comprises:

estimating a displacement of the initial static gaze point using the head pose estimation model so as to devise a final gaze point of the user.

12. An eye gaze tracking system comprising:

a detector arranged to detect a position of an iris center and a position of a corresponding eye corner of a user; and

one or more processors arranged to

determine an eye vector associated with the a gaze direction of the user using the detected iris center position and the corresponding eye corner position;

determine an initial static gaze point of the user by mapping the eye vector with a gaze target; and

apply the initial static gaze point to a head pose estimation model that is arranged to model a head pose of the user so as to devise a final gaze point of the user;

wherein the eye vector g is defined by the position of the iris center p_iris and the position of the eye corner p_corner via a relationship of g=p_corner−p_iris; and

wherein the head pose estimation model is arranged to apply a sinusoidal head model to simulate a shape of the head of the user.

13. An eye gaze tracking system in accordance with claim 12 , wherein the detector and the one or more processors are further arranged

to detect and extract an eye region from an image of the user; and

to detect the position of the iris center and the position of the corresponding eye corner from the extracted eye region.

14. An eye gaze tracking system in accordance with claim 12 , further comprising a display arranged to present a gaze target for view by the user; and wherein the detector and the one or more processors are arranged to determine an eye vector of the user for viewing the gaze target so as to map the eye vector with a gaze target.

15. A non-transient computer readable medium for storing computer instructions that, when executed by at least one processor, causes the at least one processor to perform an eye gaze tracking method comprising the steps of:

detecting a position of an iris center and a position of a corresponding eye corner of a user;

determining an eye vector associated with the a gaze direction of the user using the detected iris center position and the corresponding eye corner position;

determining an initial static gaze point of the user by mapping the eye vector with a gaze target; and

applying the initial static gaze point to a head pose estimation model that is arranged to model a head pose of the user so as to devise a final gaze point of the user;

wherein the eye vector g is defined by the position of the iris center p_iris and the position of the eye corner p_corner via a relationship of g=p_corner−p_iris; and

wherein the head pose estimation model is arranged to apply a sinusoidal head model to simulate a shape of the head of the user.