User-Calibration-Free Gaze Estimation Method Using a Binocular 3D Eye Model

Takashi NAGAMATSU; Ryuichi SUGANO; Yukina IWAMOTO; Junzo KAMAHARA; Naoki TANAKA

doi:10.1587/transinf.E94.D.1817

User-Calibration-Free Gaze Estimation Method Using a Binocular 3D Eye Model

Takashi NAGAMATSU, Ryuichi SUGANO, Yukina IWAMOTO, Junzo KAMAHARA, Naoki TANAKA

Full Text Views

0

Cite this

Summary :

This paper presents a user-calibration-free method for estimating the point of gaze (POG). This method provides a fast and stable solution for realizing user-calibration-free gaze estimation more accurately than the conventional method that uses the optical axis of the eye as an approximation of the visual axis of the eye. The optical axis of the eye can be estimated by using two cameras and two light sources. This estimation is carried out by using a spherical model of the cornea. The point of intersection of the optical axis of the eye with the object that the user gazes at is termed POA. On the basis of an assumption that the visual axes of both eyes intersect on the object, the POG is approximately estimated using the binocular 3D eye model as the midpoint of the line joining the POAs of both eyes. Based on this method, we have developed a prototype system that comprises a 19^″ display with two pairs of stereo cameras. We evaluated the system experimentally with 20 subjects who were at a distance of 600 mm from the display. The root-mean-square error (RMSE) of measurement of POG in the display screen coordinate system is 1.58.

Publication: IEICE TRANSACTIONS on Information Vol.E94-D No.9 pp.1817-1829

Publication Date: 2011/09/01

Publicized

Online ISSN: 1745-1361

DOI: 10.1587/transinf.E94.D.1817

Type of Manuscript: PAPER

Category: Multimedia Pattern Processing

Cite this

Copy

Takashi NAGAMATSU, Ryuichi SUGANO, Yukina IWAMOTO, Junzo KAMAHARA, Naoki TANAKA, "User-Calibration-Free Gaze Estimation Method Using a Binocular 3D Eye Model" in IEICE TRANSACTIONS on Information, vol. E94-D, no. 9, pp. 1817-1829, September 2011, doi: 10.1587/transinf.E94.D.1817.
Abstract: This paper presents a user-calibration-free method for estimating the point of gaze (POG). This method provides a fast and stable solution for realizing user-calibration-free gaze estimation more accurately than the conventional method that uses the optical axis of the eye as an approximation of the visual axis of the eye. The optical axis of the eye can be estimated by using two cameras and two light sources. This estimation is carried out by using a spherical model of the cornea. The point of intersection of the optical axis of the eye with the object that the user gazes at is termed POA. On the basis of an assumption that the visual axes of both eyes intersect on the object, the POG is approximately estimated using the binocular 3D eye model as the midpoint of the line joining the POAs of both eyes. Based on this method, we have developed a prototype system that comprises a 19^″ display with two pairs of stereo cameras. We evaluated the system experimentally with 20 subjects who were at a distance of 600 mm from the display. The root-mean-square error (RMSE) of measurement of POG in the display screen coordinate system is 1.58.
URL: https://global.ieice.org/en_transactions/information/10.1587/transinf.E94.D.1817/_p

Copy

@ARTICLE{e94-d_9_1817,
author={Takashi NAGAMATSU, Ryuichi SUGANO, Yukina IWAMOTO, Junzo KAMAHARA, Naoki TANAKA, },
journal={IEICE TRANSACTIONS on Information},
title={User-Calibration-Free Gaze Estimation Method Using a Binocular 3D Eye Model},
year={2011},
volume={E94-D},
number={9},
pages={1817-1829},
abstract={This paper presents a user-calibration-free method for estimating the point of gaze (POG). This method provides a fast and stable solution for realizing user-calibration-free gaze estimation more accurately than the conventional method that uses the optical axis of the eye as an approximation of the visual axis of the eye. The optical axis of the eye can be estimated by using two cameras and two light sources. This estimation is carried out by using a spherical model of the cornea. The point of intersection of the optical axis of the eye with the object that the user gazes at is termed POA. On the basis of an assumption that the visual axes of both eyes intersect on the object, the POG is approximately estimated using the binocular 3D eye model as the midpoint of the line joining the POAs of both eyes. Based on this method, we have developed a prototype system that comprises a 19^″ display with two pairs of stereo cameras. We evaluated the system experimentally with 20 subjects who were at a distance of 600 mm from the display. The root-mean-square error (RMSE) of measurement of POG in the display screen coordinate system is 1.58.},
keywords={},
doi={10.1587/transinf.E94.D.1817},
ISSN={1745-1361},
month={September},}

Copy

TY - JOUR
TI - User-Calibration-Free Gaze Estimation Method Using a Binocular 3D Eye Model
T2 - IEICE TRANSACTIONS on Information
SP - 1817
EP - 1829
AU - Takashi NAGAMATSU
AU - Ryuichi SUGANO
AU - Yukina IWAMOTO
AU - Junzo KAMAHARA
AU - Naoki TANAKA
PY - 2011
DO - 10.1587/transinf.E94.D.1817
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E94-D
IS - 9
JA - IEICE TRANSACTIONS on Information
Y1 - September 2011
AB - This paper presents a user-calibration-free method for estimating the point of gaze (POG). This method provides a fast and stable solution for realizing user-calibration-free gaze estimation more accurately than the conventional method that uses the optical axis of the eye as an approximation of the visual axis of the eye. The optical axis of the eye can be estimated by using two cameras and two light sources. This estimation is carried out by using a spherical model of the cornea. The point of intersection of the optical axis of the eye with the object that the user gazes at is termed POA. On the basis of an assumption that the visual axes of both eyes intersect on the object, the POG is approximately estimated using the binocular 3D eye model as the midpoint of the line joining the POAs of both eyes. Based on this method, we have developed a prototype system that comprises a 19^″ display with two pairs of stereo cameras. We evaluated the system experimentally with 20 subjects who were at a distance of 600 mm from the display. The root-mean-square error (RMSE) of measurement of POG in the display screen coordinate system is 1.58.
ER -