SYSTEM AND METHOD OF DETECTING EYE FIXATIONS USING ADAPTIVE THRESHOLDS

• US 20090086165A1
• Filed: 09/28/2007
• Published: 04/02/2009
• Est. Priority Date: 09/28/2007
• Status: Active Grant

• Current Assignee: Google LLC (Alphabet Inc.)
• Original Assignee: International Business Machines Corporation
• Inventors: BEYMER, David James
• Granted Patent: US 7,556,377 B2
• Time in Patent Office: Days
• US Class Current: 351/210
• CPC Class Codes: A61B 3/113 for determining or recording eye movement
• CPC Class Codes: G06V 40/19 Sensors therefor

Abstract

A system and method of adaptively establishing fixation thresholds for eye-gaze tracking data and identifying fixations within the eye-gaze tracking data are disclosed. Eye-gaze tracking data is collected. A saccade estimate may be calculated using a percentile statistic of the changes in eye-gaze position. Fixations may be determined by comparing the saccade estimates with the changes in eye-gaze positions over time windows.

Claims

1. An adaptive eye-tracking system comprising

a visual display having display points represented by display point coordinates and display data including display point coordinates;

a camera generating video data representing a user'"'"'s eye viewing the visual display; and

a processor receiving said video data from said camera, estimating gaze data indicating eye-gaze positions and including time stamps, and associating said eye gaze positions with display point coordinates, calculating a gaze noise value using said eye gaze positions, calculating a saccade estimate using said gaze noise value and determining display point coordinates representing fixations using said saccade estimate and said associated eye gaze positions.

2. The adaptive eye-tracking system of claim 1, wherein said gaze noise value is calculated using changes in eye gaze positions over a period of time.

3. The adaptive eye-tracking system of claim 2, wherein said gaze noise value is calculated as a selected percentile statistic of the changes in eye gaze positions over a period of time.

4. The adaptive eye-tracking system of claim 1, wherein display point coordinates representing fixations are determined by comparing a change in eye position over a time window with the saccade estimate.

5. The adaptive eye-tracking system of claim 1, wherein said display point coordinates include horizontal and vertical coordinates.

6. The adaptive eye-tracking system of claim 5, wherein a horizontal saccade estimate is calculated and a vertical saccade estimate is calculated.

7. The adaptive eye-tracking system of claim 6 wherein display point coordinates representing fixations are determined by comparing a horizontal dispersion in eye position with a horizontal saccade estimate and by comparing a vertical dispersion in eye position with a vertical saccade estimate.

8. An adaptive eye-tracking method comprising:

recording eye-gaze positions;

calculating a saccade estimate using said recorded eye-gaze positions;

determining eye-gaze fixations using said recorded eye-gaze positions and said saccade estimate.

9. The adaptive eye-tracking method of claim 8 where the said recorded eye-gaze positions consist of horizontal and vertical coordinates.

10. The adaptive eye-tracking method of claim 8 further comprising calculating changes in eye-gaze positions.

11. The adaptive eye-tracking method of claim 10, wherein said determining eye-gaze fixations comprises comparing the dispersion in eye-gaze positions with said saccade estimate.

12. The adaptive eye-tracking method of claim 10, wherein said saccade estimate is calculated using changes in eye-gaze positions.

13. The adaptive eye-tracking method of claim 10, wherein said saccade estimate is calculated using a percentile statistic of the changes in eye-gaze positions.

14. The adaptive eye-tracking method of claim 8, further comprising associating said eye-gaze fixations with points on a display.

15. The adaptive eye-tracking method of claim 8, further comprising calibrating said eye-gaze positions such that each eye-gaze position of the recorded eye-gaze positions may be associated with a point on a display.

16. The adaptive eye-tracking process of claim 9, further comprising calculating a horizontal saccade estimate and calculating a vertical saccade estimate.

17. A computer program product for use with an eye-gaze tracking system, the computer program product comprising a computer useable medium including a computer readable program, wherein the computer readable program when executed on a computer causes the computer to:

generate eye-gaze tracking data, said eye-gaze tracking data including data points defined by a first coordinate value, a second coordinate value and a time-stamp value;

calculate a first coordinate percentile statistic from said first coordinate values of said data points;

calculate a first coordinate saccade estimate from said first coordinate percentile statistic;

calculate a second coordinate percentile statistic from said second coordinate values of said data points;

calculate a second coordinate saccade estimate from said second coordinate percentile statistic;

determine the presence of a fixation in a series of data points having sequential time-stamp values by comparing dispersion in first coordinate values with said first coordinate saccade estimate and comparing dispersion in second coordinate values with said second coordinate saccade estimate; and

indicate the first coordinate values, second coordinate values, and beginning and ending time-stamp values of said fixation.

18. The computer program product of claim 17, wherein said first coordinate value corresponds to a vertical position on a display.

19. The computer program product of claim 17, further comprising receiving eye-gaze data from a camera, wherein said eye-gaze tracking data is generated from said eye-gaze data.