Implicit Gaze Calibration Using Forward-Axis Eye Tracking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing head-up display systems face challenges in accurately estimating optical axes and focal depths due to deviations in human eye optical axes, limitations in head pose estimation, and inaccuracies in gaze and focal depth estimation, necessitating inconvenient explicit calibration processes that disrupt user experience.
Innovation Solution
A system and method for implicit gaze and focus distance calibration using a tracking camera to capture images, determine tracking parameters, estimate uncalibrated gaze vectors, and calculate interpupillary distance based on predefined forward axis of vision, eliminating the need for explicit user calibration.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If explicit gaze calibration process is implemented, then gaze estimation accuracy is improved, but user convenience and viewing experience deteriorate
Solution Approach 1:
The system performs automatic calibration without requiring user participation. The processor captures images, detects eye features, estimates optical axes and focal depths, and calibrates gaze vectors autonomously, allowing the system to serve itself rather than requiring user action for calibration
Solution Approach 2:
The calibration process is performed automatically in the background before the actual gaze tracking task begins. The system pre-calibrates the gaze vectors by capturing multiple images, detecting eye features, and computing calibration parameters without interrupting the user's primary task
2Measurement precision
If explicit gaze calibration process is implemented, then gaze estimation accuracy is improved, but time consumption increases
Solution Approach 1:
The system performs automatic calibration without requiring user participation. The processor captures images, detects eye features, estimates optical axes and focal depths, and calibrates gaze vectors autonomously, allowing the system to serve itself rather than requiring user action for calibration
Solution Approach 2:
The calibration process is seamlessly integrated into the system operation without interrupting the user's primary task. The automatic calibration occurs in the background, maintaining continuous useful action while achieving accurate gaze estimation
3Device complexity
If pre-recorded calibration databases are used, then system complexity is reduced, but calibration reliability deteriorates
Solution Approach 1:
The system computes individualized calibration parameters for each user by capturing multiple images, detecting eye features, and calculating optical axes and focal depths specific to that user's anatomy. This dynamic parameter adjustment based on individual user characteristics improves reliability over generic pre-recorded databases
Data Source
AI summary
Images of a user's face are captured at a plurality of time instants. Tracking parameters are determined, tracking parameters include: a pose of the user's head, positions of eyeballs, and at least one of: relative positions of irises with respect to boundaries of the eyeballs, relative positions of irises with respect to corners, shapes of the user's eyes. Uncalibrated gaze vectors of the user's eyes are estimated. A set of uncalibrated gaze vectors is generated. The uncalibrated gaze vectors of thset are stored along with corresponding time instants and tracking parameters. A first subset of uncalibrated gaze vectors whose direction matches with a predefined forward axis of vision, is selected. For the uncalibrated gaze vectors, corresponding tracking parameters are fetched. A maximum distance between the irises is determined. The maximum distance is considered as an interpupillary distance.