Eye Tracking Using Calibrated Viewpoint Matrix
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current eye tracking methods in VR/AR systems face challenges in accurately detecting and tracking gaze directions, especially in head-mounted displays, due to variations in pupil shape and head movements, leading to reduced accuracy and user experience.
Innovation Solution
An electronic device with a camera and processing circuit that captures eye images, detects pupil regions of interest, analyzes gaze vectors using an eye model with a matrix indicating the relationship between the viewpoint and gaze vector, and tracks eye motion to improve accuracy and user interaction in VR/AR environments.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional eye tracking methods are used in HMD, then the system can detect gaze direction, but the accuracy is reduced due to variations in pupil shape and head movements
Solution Approach 1:
The patent applies parameter changes by using an eye model with a matrix that represents the relationship between viewpoint and gaze vector. This matrix is calibrated to account for individual variations in pupil shape and head movement characteristics, transforming the detection parameters to maintain high accuracy across different users and conditions.
Solution Approach 2:
The patent implements preliminary action through the calibration process that establishes the eye model matrix before actual gaze tracking. During calibration, the system pre-computes the relationship between viewpoint and gaze vector for the specific user, storing this information for use during operation. This preliminary setup enables accurate tracking without requiring real-time adjustments for individual variations.
2Ease of operation
If eye tracking is implemented in HMD, then user interaction can be enhanced, but the complexity of the system increases due to calibration requirements
Solution Approach 1:
The patent applies self-service by designing a calibration process that the user performs themselves during initial setup. The system guides the user through calibration steps where they simply need to look at different positions, and the system automatically computes the eye model matrix. This eliminates the need for complex manual configuration or professional calibration, making the system easy to operate while maintaining accuracy.
Data Source
AI summary
An eye tracking method includes: capturing, by a camera, an image of an eye; detecting, by a processing circuit, a pupil region of interest in the image of the eye; analyzing, by the processing circuit, the pupil region of interest to obtain a gaze vector of the pupil region of interest; calculating, by the processing circuit, a viewpoint of the eye according to the gaze vector based on an eye model, in which the eye model includes a matrix indicating relationship between the viewpoint of the eye and the gaze vector of the pupil region of interest; and tracking, by the processing circuit, a motion of the eye based on the viewpoint calculated using the eye model.