Dmitri Model

An Automatic Personal Calibration Procedure for Advanced Gaze Estimation Systems

Gaze estimation systems use calibration procedures to estimate subject-specific parameters that are needed for the calculation of the point-of-gaze. In these procedures, subjects are required to fixate on a specific point or points in space at specific time instances. Advanced remote gaze estimation systems can estimate the optical axis of the eye without any personal calibration procedure, but use a single calibration point to estimate the angle between the optical axis and the visual axis (line-of-gaze). This paper presents a novel calibration procedure that does not require active user participation. To estimate the angles between the optical and visual axes of each eye, this procedure minimizes the distance between the intersections of the visual axes of the left and right eyes with one or more observation surfaces (displays) while subjects look naturally at these displays (e.g., watching a video clip). Theoretical analysis and computer simulations show that the performance of the proposed procedure improves when the range of angles between the visual axes and vectors normal to the observation surfaces increases. Experiments with four subjects show that the subject-specific angles between the optical and visual axes can be estimated with an rms error of 0.5°.

User-calibration-free remote gaze estimation system

Gaze estimation systems use calibration procedures that require active subject participation to estimate the point-of-gaze accurately. In these procedures, subjects are required to fixate on a specific point or points in space at specific time instances. This paper describes a gaze estimation system that does not use calibration procedures that require active user participation. The system estimates the optical axes of both eyes using images from a stereo pair of video cameras without a personal calibration procedure. To estimate the point-of-gaze, which lies along the visual axis, the angles between the optical and visual axes are estimated by a novel automatic procedure that minimizes the distance between the intersections of the visual axes of the left and right eyes with the surface of a display while subjects look naturally at the display (e.g., watching a video clip). Experiments with four subjects demonstrate that the RMS error of this point-of-gaze estimation system is 1.3°.

An Automated Hirschberg Test for Infants

A novel automated method to measure eye misalignment in infants is presented. The method uses estimates of the Hirschberg ratio (HR) and angle Kappa (the angle between the visual and optical axis) for each infant to calculate the angle of eye misalignment. The HR and angle Kappa are estimated automatically from measurements of the direction of the optical axis and the coordinates of the center of the entrance pupil and corneal reflexes in each eye when infants look at a set of images that are presented sequentially on a computer monitor. The HR is determined by the slope of the line that describes the direction of the optical axis as a function of the distance between the center of the entrance pupil and the corneal reflexes. The peak of the distribution of possible angles Kappa during the image presentation determines the value of angle Kappa. Experiments with five infants showed that the 95% limits of agreement between repeated measurements of angle Kappa are 0.61. The maximum error in the estimation of eye alignment in orthotropic infants was 0.9 with 95% limits of agreement between repeated measurements of 0.75.

Fixation-free assessment of the Hirschberg ratio.

PURPOSE To describe a novel methodology by which to measure the Hirschberg ratio (HR) in infants. The methodology does not require fixation on specific points, and measurements are made while infants look naturally at a display. METHODS The HR is calculated automatically from measurements of the direction of the optical axis, the position of the pupil center, and corneal reflexes in video images from an advanced two-camera eye-tracking system. The performance of the novel fixation-free procedure (FFP) was evaluated in 43 adults by measuring the average difference and the 95% limits of agreement with the standard fixation-based procedure (FBP). Repeatability of the HR measurements was evaluated by assessing the 95% limits of agreement between two independent measurements. Performance of the FFP was also evaluated in five infants. RESULTS In adults, the average HR was 12.89 +/- 1.22 degrees/mm for FFP and 12.81 +/- 1.22 degrees/mm for FBP. FFP and FBP measurements were highly correlated (r = 0.95; P < 0.001). The 95% limits of agreement between FFP and FBP were +/-0.86 degrees/mm. The 95% limits of agreement of repeated measurements were +/-0.66 degrees/mm for FFP and +/-0.77 degrees/mm for FBP. In infants, the 95% limits of agreement of repeated measurements by FFP were +/-0.63 degrees/mm. CONCLUSIONS In adults, the FFP provides accurate measurements of the HR that are in excellent agreement with measurements by FBP. In infants, measurements of HR by FFP show the same repeatability and consistency.

A probabilistic approach for the estimation of angle kappa in infants

This paper presents a probabilistic approach for the estimation of the angle between the optical and visual axes (angle kappa) in infants. The approach assumes that when patterned calibration targets are presented on a uniform background, subjects are more likely to look at the calibration targets than at the uniform background, but it does not require accurate and continuous fixation on presented targets. Simulations results show that when subjects attend to roughly half of the presented targets, angle kappa can be estimated accurately with low probability (< 1%) of false detection. In experiments with five babies who attended to the calibration target for only 47% of the time (range from 26% to 70%), the average difference between repeated measurements of angle kappa was 0.04 ± 0.31°.

An automatic calibration procedure for remote eye-gaze tracking systems

Remote gaze estimation systems use calibration procedures to estimate subject-specific parameters that are needed for the calculation of the point-of-gaze. In these procedures, subjects are required to fixate on a specific point or points at specific time instances. Advanced remote gaze estimation systems can estimate the optical axis of the eye without any personal calibration procedure, but use a single calibration point to estimate the angle between the optical axis and the visual axis (line-of-sight). This paper presents a novel automatic calibration procedure that does not require active user participation. To estimate the angles between the optical and visual axes of each eye, this procedure minimizes the distance between the intersections of the visual axes of the left and right eyes with the surface of a display while subjects look naturally at the display (e.g., watching a video clip). Simulation results demonstrate that the performance of the algorithm improves as the range of viewing angles increases. For a subject sitting 75 cm in front of an 80 cm × 60 cm display (40” TV) the standard deviation of the error in the estimation of the angles between the optical and visual axes is 0.5°.

Covert monitoring of the point-of-gaze

Gaze estimation systems use calibration procedures that require active subject participation to estimate the point-of-gaze accurately. Consequently, these systems do not support covert monitoring of visual scanning patterns. This paper presents a novel gaze estimation methodology that does not use calibration procedures that require active user participation. This methodology uses multiple infrared light sources for illumination and a stereo pair of video cameras to obtain images of the eyes. Each pair of images is analyzed and the centers of the pupils and the centers of curvature of the corneas are estimated. These points, which are estimated without a personal calibration procedure, define the optical axis of each eye. To estimate the point-of-gaze, which lies along the visual axis, the angle between the optical and visual axes is estimated by a procedure that minimizes the distance between the intersections of the visual axes of the left and right eyes with the surface of a display while subjects look naturally at the display (e.g., watching a video clip). Simulation results demonstrate that for a subject sitting 75 cm in front of an 80 cm × 60 cm display (40″ TV) the RMS error of the estimated point-of-gaze is 17.8 mm (1.3°).