Eye Tracking Device 3D Model Superimposition Calibration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing eye tracking devices require users to perform an image frame preparing operation during initial calibration, which burdens the user and is inefficient.
Innovation Solution
An eye tracking device that includes an image acquiring unit, an arithmetic processing unit performing image processing, and a line-of-sight angle calculating unit, utilizing a three-dimensional face model for superimposition processing to estimate the line-of-sight direction without the need for initial calibration or image frame preparation, by correcting the eyeball position based on the distance from the nose point to the pupil.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional line-of-sight direction estimating devices use two or more image frame sequences obtained by user changing face direction during initial calibration, then measurement precision of line-of-sight direction is improved, but ease of operation deteriorates due to burden on user
Solution Approach 1:
The system uses the user's own facial image captured by the camera to automatically generate the three-dimensional face model through image processing, eliminating the need for the user to manually change face directions during calibration. The user simply needs to provide one facial image, and the system automatically performs all necessary processing to create the 3D model for accurate line-of-sight estimation.
Solution Approach 2:
The patent replaces the mechanical approach of manually capturing multiple image frames by having the user change face directions with an automated image processing system. The system uses computer vision algorithms to process a single facial image and generate a three-dimensional face model, substituting manual mechanical operations with automated computational methods.
2Reliability
If initial calibration with image frame preparation is performed, then reliability of line-of-sight tracking is improved, but loss of time increases due to calibration requirement
Solution Approach 1:
The system performs preliminary action by automatically generating the three-dimensional face model from a single captured facial image during the initial setup phase. This pre-processing creates a ready-to-use 3D model that enables reliable line-of-sight tracking without requiring time-consuming calibration procedures before actual use.
3Measurement precision
If three-dimensional model superimposition processing is used to estimate eyeball position, then measurement precision of pupil direction is improved, but device complexity increases
Solution Approach 1:
The patent introduces a three-dimensional face model as an intermediary between the captured facial image and the line-of-sight direction estimation. This 3D model serves as a mediator that simplifies the complex task of estimating eyeball position and pupil direction by providing a standardized geometric representation that can be easily processed and superimposed on the original image.
Data Source
AI summary
An eye tracking device according to the present disclosed technology includes: an image acquiring unit that acquires an image of a subject; an arithmetic processing unit that performs image processing; and a line-of-sight angle calculating unit that calculates a line-of-sight direction vector on the basis of a result of the image processing. The arithmetic processing unit includes: a real image arithmetic processing unit that performs arithmetic processing on a real image in an image space; and a three-dimensional model superimposition processing unit that performs superimposition processing on the image space using a three-dimensional face model. The three-dimensional model superimposition processing unit includes a three-dimensional model correcting unit that corrects the three-dimensional face model in order to eliminate an error caused by the superimposition processing.