Gaze Tracking via Light Path Tracing and Refraction Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional gaze tracking methods face challenges in accurately determining the position and orientation of the eye due to factors like refraction through glasses and non-spherical cornea models, often relying on approximations or ignoring these factors, which can affect the accuracy of gaze point estimation.
Innovation Solution
A method that estimates preliminary light paths based on the eye's position and orientation, forming distances between detected and predicted features, and uses an objective function to optimize the determination of the eye's position and orientation, taking into account refraction and non-spherical cornea models, to improve gaze tracking accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If traditional gaze tracking methods use simple approximations or ignore refraction and non-spherical cornea models, then the device complexity is reduced, but the measurement precision of gaze point estimation deteriorates
Solution Approach 1:
The patent applies parameter changes by transitioning from simplified eye models to a more accurate model that incorporates refraction effects and non-spherical cornea geometry. The method changes the parameters of the optical model to match real eye anatomy, including corneal curvature variations and refractive indices, thereby improving measurement precision without requiring complex additional hardware
Solution Approach 2:
The patent replaces simple geometric optics calculations with a more sophisticated optical model that accounts for refraction through the cornea and aqueous humor. This substitution uses ray tracing algorithms that incorporate Snell's law and non-spherical surface geometry, trading increased computational complexity for significantly improved gaze estimation accuracy
2Speed
If traditional gaze tracking methods use simple approximations, then the computational speed is improved, but the reliability of gaze tracking deteriorates
Solution Approach 1:
The patent applies preliminary action by pre-calculating and storing refraction correction parameters and corneal surface characteristics during a calibration phase. These pre-computed parameters are then reused during actual gaze tracking, reducing the computational burden in real-time operation while maintaining high reliability through accurate optical modeling
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach enhances the accuracy of gaze tracking by effectively handling refraction and non-spherical cornea models, providing a more precise estimation of the eye's position and orientation, thereby improving the reliability of gaze point determination.
Implementation Method 1
The image captured by the camera shows a first reflection of the illuminator at a cornea of the eye
Implementation Method 2
estimating, based on the preliminary position of the eye and the preliminary orientation of the eye, a second preliminary path for light travelling through at least a portion of the cornea of the eye towards the camera
Data Source
Figure 1~2
Figure 3~4
Figure 5~6
AI summary
A preliminary path (410) for light travelling towards a camera (302) via corneal reflection (116) is estimated based on a preliminary position (109) and orientation (112) of an eye (100). A position (502) where the reflection would appear in images (500) captured by the camera is estimated. A distance (503) is formed between a detected position (501) of a corneal reflection of an illuminator (301) and the estimated position. A second preliminary path (810) for light travelling through the cornea (106) or from the sclera (117) towards a camera (302) is estimated based on the preliminary position and orientation, and a position (902, 905) where the second preliminary path would appear to originate in images (900) captured by this camera is estimated. A distance (903, 906) is formed between a detected edge (901, 904) of a pupil (103) or iris (101) and the estimated position where the second preliminary path would appear to originate. An updated position and/or orientation of the eye is determined using an objective function formed based on the formed distances.