FEATURES OF THE EYEGAZE SYSTEM

Calibration

The procedure to calibrate the Eyegaze System is robust yet fast and easy to perform. The calibration procedure takes approximately 15 seconds.

Fully Automatic Operation: The calibration procedure is fully automatic; no assistance from another person is required. The procedure adapts to the user speed by waiting for the user to fixate clearly on each calibration point before accepting it and moving on to the next point. The procedure accommodates interruptions from the user blinking or looking away from the computer screen. The procedure simply waits for a good fixation before moving to the next calibration point.

Consistency Checking: After the original pass through the calibration points, the procedure tests that the eye was properly fixated on each point by checking that each gazepoint prediction is consistent with all the other calibration points. It retakes any calibration points that are inconsistent with other points. The procedure does not accept the full calibration until the overall gaze prediction accuracy and consistency exceed desired thresholds.

Accuracy

To achieve high gazepoint tracking accuracy, the image processing algorithms in the Eyegaze System explicitly accommodate several common sources of gazepoint tracking error.

Nonlinear Gazepoint Tracking Equations: The Eyegaze System uses nonlinear equations to predict the gazepoint accurately from the measured glint-pupil vector. The nonlinear terms account for a) the camera axis being tilted with respect to the computer screen, b) curved computer display screens, and c) flattening of the corneal surface toward the edges.

Accommodating Head Range Variation: The accuracy of video eyetrackers are typically sensitive to head motion along the camera axis. As the head moves toward the camera the predicted gaze point (if uncorrected for range) moves radially away from the camera, and as the head moves backward, the predicted gaze point moves radially in toward the camera. Typically, when a person is about 24 inches from the camera, and looking at a point toward the top of the computer screen head motions of 1.0 inch along the camera Z axis result in predicted gazepoint variations of about 0.75 inch.

The Eyegaze System uses the patented Asymmetric Aperture Method to measure variations in the range between the camera and the cornea of the eye, and it uses the range information to minimize gazepoint tracking errors resulting from longitudinal head motions.

Accommodating Pupil Diameter Variation: With video eyetrackers, the center of the pupil is not directly measurable from the camera's image of the eye. The pupil center is estimated by observing the edges of the pupil and calculating the center location from the edge measurements. Due to the fact that the pupil lies behind the corneal surface of the eye, however, a ray from the center of the physical pupil does not arrive precisely at the center of the pupil image. When the eye is looking away from the camera, the curved cornea refracts the rays from the various pupil edge points differently. Thus as pupil diameter varies concentrically about its true center, the edges in the pupil image move nonconcentrically around the true pupil center point, even if the true pupil center is stationary.

If varying pupil diameter is not explicitly accommodated, gazepoint calculations can vary up to 0.5 inch as the pupil diameter varies between 3.5 and 7 mm. The Eyegaze System has corneal refraction compensation logic to minimize gazepoint calculation errors resulting from varying pupil diameter.

Accommodating Glint Straddling Pupil Edge: Many PCCR eyetrackers are prone to errors when the corneal reflection is near or straddles the edge of the pupil. Pupil edge measurements in the region of the corneal reflection are unreliable, and image intensity gradients from the pupil-iris boundary distort the corneal reflection image and result in imprecise measurement of the true corneal reflection location. The Eyegaze System has sophisticated image processing software to locate the corneal reflection accurately when the corneal reflection straddles the pupil-iris boundary.

Reliability

Reliability refers to the range of operational conditions under which the Eyegaze System is able to measure gazepoint data.

Robust Eye Recognition: Rather than using simple brightness thresholding methods to detect the pupil and corneal reflection, the Eyegaze System uses sophisticated pattern recognition and hypothesis testing logic to detect the eye reliably in cluttered images. The system does not get confused by reflections off glasses, or by bright or dark facial features. The system can not measure gaze direction if reflection from glasses is directly superimposed on the pupil and/or corneal reflection, but it is not confused by these spurious reflections. It simply reports that a "measurable eye" was not found.

Accommodating Human Eye Variations: The system accommodates a wide variation in the eye image, such as pupil size and pupil and iris brightness and it accommodates to most glasses and contacts. The system typically tracks 90 percent of the human population.