Eyeball Model And Gaze Estimation Method Research

The eye is one of the most important seilse organs of human beings.About 80 percent to 90 percent of humans' information received from the outside world comes from vision.With the development of science and technology,the observation of the eyes has attracted the attention of a large number of researchers.And more and more scientific workers have invested in the research of gaze estimation technology.Gaze estimation technology has a large number of applications in various fields such as user experience,interactive research,marketing research,consumer research,psychology and neuroscience,and even security and law enforcement.Firstly,this paper introduces the research background and development history of gaze method,and summarizes the research status of gaze method at home and abroad,and also briefly introduces the application of gaze estimation method.Understanding the eyeball structure is the premise of studying gaze estimation method based on the three-dimensional eyeball model.So the eyeball is also described in detail in terms of its physiological structure,motion law,and the like.At the same time,an in-depth study was made on the method of line-of-sight estimation based on the three-dimensional eyeball model.For the three different three-dimensional models based on the spherical eyeball model,the rotational symmetry eyeball model and the quadric surface eyeball model,three different solving methods are proposed and corresponding simulation experiments are performed.In the spherical eyeball model,the line of sight estimation mainly uses the phenomenon of refraction and reflection when incident light is on the surface of the cornea of the eyeball.The geometrical model of reflection and refraction is used to solve the center of the cornea and the center of the pupil,thereby reconstructing the line of sight of the eyeball.In a rotationally symmetric eyeball model,because the backward extension line of the reflected light on the corneal surface that does not converge on the particularity of the corneal curvature center,special settings need to be made on the experimental equipment,that is,the infrared light source and the camera are placed on the same position,then use the coplanar relationship to solve the optical axis.And then the co-planar relationship is used to solve the optical axis.Finally,using a simulated eye tracker data framework.This experimental conditions were set and eye movement data were obtained under corresponding conditions.These data were used to verify the Correctness of the rotational symmetry model.Another eye-sight estimation method is based on a three-dimensional eye model of a ellipsoid surface model.Although the use of the rotating model in Chapter 3 has become more relevant to the actual human eye.However,this method still only defines the solution process of the optical axis as an aspherical solution.The solution of the curvature center of the actual cornea is not based on the aspherical 3D eye.Therefore,in order to more closely match the actual human eye,the accuracy of the line-of-sight estimation method is further enhanced.And this method's application range will be further expanded.Using the new ellipsoid surface eyeball model,the optical properties of the quadratic surface normal and the reflection and refraction relationships are used to solve the optical axis of the eyeball model and the corneal curvature center.The experimental results show that the average error of the corneal curvature radius solved by this method is only 0.00634 mm.