Gaze Tracking Using Eye White Pixel Ratios
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current gaze tracking technologies in entertainment environments are limited to short distances (up to three feet) and fail to accurately detect gaze in varying lighting conditions, making them impractical for interactions with animatronics or other interactive devices in theme parks and similar settings.
Innovation Solution
A computer program and system that captures images of participants, calculates the ratio of white pixels to iris pixels in both eyes to estimate gaze direction, allowing for accurate gaze detection at distances up to twelve feet and in varying lighting conditions, using inexpensive image capture devices.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Length of stationary object
If conventional facial tracking software is used, then gaze direction can be determined, but the tracking distance is limited to approximately two to three feet
Solution Approach 1:
The patent transitions from 2D facial feature detection to 3D spatial analysis by introducing depth information through stereo vision. The system uses multiple cameras to capture images from different angles and reconstructs three-dimensional eye position and gaze direction, enabling accurate tracking at distances beyond the conventional two to three feet limitation while maintaining measurement precision through spatial triangulation
Solution Approach 2:
The patent creates a multi-functional gaze tracking system that can operate across multiple distance ranges (from close-up to twelve feet or more) using the same hardware configuration. The system adapts its processing algorithms based on detected distance, switching between different analysis methods to maintain accuracy across varying spatial scales, making it universally applicable for both close-range and far-range gaze tracking
2Adaptability or versatility
If conventional facial tracking software is used, then gaze can be tracked, but it requires very particular lighting conditions
Solution Approach 1:
The patent employs adaptive parameter adjustment that dynamically modifies processing thresholds, contrast sensitivity, and feature detection criteria based on the detected lighting conditions. The system automatically calibrates its analysis parameters in response to varying illumination levels, color temperatures, and light directions, maintaining reliable gaze tracking accuracy across diverse lighting environments from dimly lit to brightly lit spaces
Solution Approach 2:
The system implements dynamic adaptation to changing lighting conditions by continuously monitoring environmental light characteristics and adjusting its processing pipeline in real-time. The algorithm dynamically reweights different visual features based on current lighting quality, enhancing robustness against lighting variations and maintaining tracking reliability without requiring fixed or controlled illumination
3Measurement precision
If Fraunhofer face tracking configuration is used, then face tracking is achieved, but the cameras are relatively much more expensive than a webcam
Solution Approach 1:
The patent replaces expensive specialized gaze-tracking cameras with inexpensive standard webcams or consumer-grade image sensors. By leveraging advanced image processing algorithms, the system achieves professional-grade gaze detection accuracy using commodity hardware that is readily available and cost-effective, dramatically reducing system manufacturing costs while maintaining measurement precision
Solution Approach 2:
The patent substitutes specialized optical-mechanical gaze tracking hardware with a software-based processing system that runs on standard imaging devices. Instead of relying on expensive dedicated sensors and complex mechanical eye-tracking apparatus, the invention uses computational methods to extract gaze information from ordinary camera images, replacing mechanical complexity with algorithmic intelligence
Data Source
AI summary
An image of a person is captured. A left eye image is located in the image. A right eye image is located in the image. A first quantity of left eye white pixels and a second quantity of left eye white pixels in the left eye image locate are determined. A left eye image ratio of the first quantity of left eye white pixels to the second quantity of left eye white pixels is calculated. A first quantity of right eye white pixels and a second quantity of right eye white pixels in the right eye image is calculated. A right eye image ratio of the first quantity of right eye white pixels to the second quantity of right eye white pixels is calculated. A gaze direction of the person is determined based upon an average ratio.