System and Method For Calibrating Eye Gaze Data

• US 20140320397A1
• Filed: 04/28/2014
• Published: 10/30/2014
• Est. Priority Date: 10/27/2011
• Status: Active Grant

• Current Assignee: Mirametrix Inc. (Lattice Semiconductor Corporation)
• Original Assignee: Mirametrix Inc. (Lattice Semiconductor Corporation)
• Inventors: HENNESSEY, Craig A., FISET, Jacob, SULLIVAN, Nicholas
• Granted Patent: US 9,811,158 B2
• Time in Patent Office: Days
• US Class Current: 345/156
• CPC Class Codes: A61B 3/113 for determining or recording eye movement
• CPC Class Codes: G01S 3/7803 Means for monitoring or calibrating
• CPC Class Codes: G06F 3/013 Eye tracking input arrangements G06F3/015 takes precedence
• CPC Class Codes: G06V 40/19 Sensors therefor

Abstract

A system and method are provided for calibrating an eye gaze tracking system. The method comprises obtaining gaze data; obtaining at least one key point corresponding to a portion of media content being displayed; linking the gaze data to the at least one key point; and generating one or more calibration parameters by comparing gaze data with associated ones of the at least one key point.

Claims

1. A method of calibrating an eye gaze tracking system, the method comprising:

obtaining gaze data;

obtaining at least one key point corresponding to a portion of media content being displayed;

linking the gaze data to the at least one key point; and

generating one or more calibration parameters by comparing gaze data with associated ones of the at least one key point.

2. The method of claim 1, further comprising applying the one or more calibration parameters to the gaze data to generate calibrated gaze data;

and providing the calibrated gaze data as an output.

3. The method of claim 1, further comprising updating a calibration profile using the one or more calibration parameters.

4. The method of claim 1, wherein the obtained gaze data has been pre-calibrated using an initial or default calibration.

5. The method of claim 1, wherein each key point is represented using a key record.

6. The method of claim 5, wherein the key record comprises a key point identifier, a timestamp, a position, and a size.

7. The method of claim 1, wherein the key point is pre-associated with the media content.

8. The method of claim 1, wherein the key point is generated according to the media content

9. The method of claim 8, wherein the key point is generated by searching for a commonality of points of gaze of different users to find fixation points common to at least some of the users.

10. The method of claim 9, wherein the fixation points are filtered to find key points based on the variance and number of users associated with a particular fixation point.

11. The method of claim 8, wherein the key point is generated using any one or more of feature tracking, motion tracking, and change in brightness.

12. The method of claim 11, wherein feature tracking comprises tracking any one or more of a face, eyes, and a mouth.

13. The method of claim 11, wherein the motion tracking comprises tracking one or more fast moving objects.

14. The method of claim 1, wherein the key point is generated by modifying the media content.

15. The method of claim 1, wherein the key point is determined by detecting a user interaction.

16. The method of claim 15, wherein the user interaction comprises selection on a display.

17. The method of claim 1, wherein linking the gaze data to the at least one key point comprises determining a time lag between a computed point of gaze and an associated key point and applying the time lag to align the gaze data with the at least one key point.

18. The method of claim 1, wherein generating the one or more calibration parameters comprises obtaining a plurality of key points, selecting a subset of the plurality of key points, and determining the accuracy of computed points of gaze for each key point.

19. The method of claim 1, wherein the at least one key point is generated by:

recording points of gaze while displaying media content;

calculating potential key points;

using a measure of reliability to determine a filtered set of key points; and

using a spatial distribution model to select key points to be used.

20. The method of claim 19, wherein at least one key point is filtered based on knowledge of a direction of gaze calculated using an initial calibration.

21. The method of claim 1, wherein a calibration is determined via a linear mapping of an initial calibration to a final calibration, using key points to determine scaling and offset differences.

22. The method of claim 1, wherein a calibration is determined, without previous knowledge of x and y coordinates of points on the screen, by:

calibrating a user using a plurality of combinations of random raw gaze data point and one or more arbitrarily distributed points on the screen;

keeping at least one calibration that provides calibrated data fitting a viewing pattern of a known stimulus; and

repeating the method iteratively using at least one new arbitrarily distributed point around points paired with each raw gaze data.

23. A computer readable storage medium comprising computer executable instructions for calibrating an eye gaze tracking system, the computer executable instructions comprising instructions for performing the method of claim 1.

24. A system comprising a processor and memory, the memory comprising computer executable instructions for calibrating an eye gaze tracking system, the computer executable instructions comprising instructions for performing the method of claim 1.