3D Point-of-Gaze Estimation Using Dual Eye Line-of-Sight Vectors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing eye-gaze tracking systems can only estimate the point-of-gaze in two dimensions, which is insufficient for interacting with three-dimensional human-machine interface environments, such as volumetric displays and parallax beam splitter displays, as they do not account for the depth and complexity of these advanced interfaces.
Innovation Solution
A method and apparatus for estimating the point-of-gaze in three dimensions by capturing image data from both eyes, determining line-of-sight vectors, and using these vectors to calculate a three-dimensional point-of-gaze, while accommodating small eye movements and head movements to improve accuracy and precision.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If two-dimensional eye-gaze tracking is used, then the system is simple and easy to implement, but it cannot accurately estimate point-of-gaze in three-dimensional environments
Solution Approach 1:
The patent extends two-dimensional eye-gaze tracking to three-dimensional space by capturing images with a three-dimensional camera array and processing them to determine three-dimensional line-of-sight vectors. This allows accurate estimation of point-of-gaze in volumetric displays and other 3D environments, resolving the contradiction between measurement precision and device complexity.
2Adaptability or versatility
If three-dimensional camera array is used, then point-of-gaze estimation in 3D environments is accurate, but the system complexity increases
Solution Approach 1:
The patent creates a universal eye-gaze tracking system that can accurately track point-of-gaze across multiple display types including two-dimensional screens, three-dimensional volumetric displays, and other advanced interfaces. The three-dimensional camera array and processing methods provide a single system that adapts to various display geometries, enhancing versatility while managing complexity.
Data Source
Figure 1~2
Figure 3~5
Figure 4
AI summary
Methods for determining a point-of-gaze (POG) of a user in three dimensions are disclosed. In particular embodiments, the methods involve: presenting a three-dimensional scene to both eyes of the user; capturing image data including both eyes of the user; estimating first and second line-of-sight (LOS) vectors in a three-dimensional coordinate system for the user's first and second eyes based on the image data; and determining the POG in the three-dimensional coordinate system using the first and second LOS vectors.