Predicting Eye Gazing Parameters via Head Motion Clustering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing eye tracking devices are expensive, fragile, intrusive, and complicated to calibrate, making them unsuitable for efficiently determining optimal ophthalmic lens designs for various activities.
Innovation Solution
A method and system for predicting eye gazing parameters without using eye tracking devices, by measuring and comparing head motion parameters of an individual with those of a group of individuals performing similar activities, and using a prediction model based on clusters of visual behaviors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If eye tracking devices are used to measure visual exploration strategy, then measurement precision is improved, but device complexity and cost increase
Solution Approach 1:
The patent creates a prediction model that copies the visual exploration patterns of reference individuals. Instead of directly measuring eye movements, the system captures head motion data from reference individuals, builds predictive models of their visual exploration strategies, and uses these models to predict eye gazing parameters for new individuals based on their head motion alone.
Solution Approach 2:
The patent replaces the mechanical/optical eye tracking system with a computational prediction system. Instead of using physical sensors to track eye movements, the system uses algorithms that predict eye gazing parameters from head motion data, substituting a complex mechanical measurement system with a computational model.
2Measurement precision
If eye tracking devices are mounted on spectacle frames, then visual exploration strategy can be measured, but ease of operation deteriorates due to calibration complexity
Solution Approach 1:
The patent captures visual exploration patterns from reference individuals during calibration and stores these as predictive models. When a new individual wears the spectacles, the system copies the relevant predictive model and applies it to predict eye gazing parameters, eliminating the need for individual calibration while maintaining measurement accuracy.
Solution Approach 2:
The patent performs calibration in advance by collecting head motion and eye gazing data from reference individuals to build predictive models. This preliminary action creates a library of visual exploration patterns that can be directly applied to new users without requiring them to undergo time-consuming calibration procedures.
3Measurement precision
If eye tracking devices are used, then eye gazing parameters can be measured directly, but cost and fragility increase
Solution Approach 1:
The patent uses head motion data as a proxy or copy for direct eye movement measurement. By capturing the relationship between head motion and eye gazing in reference individuals, the system can infer eye parameters from the more robust and less expensive head motion sensors, avoiding the use of fragile eye tracking hardware.
Solution Approach 2:
The patent replaces expensive, fragile eye tracking devices with cheaper, more robust head motion sensors. The system accepts that head motion data is less direct than eye tracking data but gains reliability through the durability and lower cost of the sensing hardware, combined with computational prediction to recover eye gazing information.
Data Source
AI summary
This method for predicting at least one eye gazing parameter of a person wearing visual equipment includes steps of: measuring data of at least two individuals wearing visual equipment, such data including at least head motion parameters and eye gazing parameters of the individuals; measuring head motion parameters of the person in real-life conditions; comparing the head motion parameters of the person with the head motion parameters of the at least two individuals; and predicting the at least one eye gazing parameter of the person at least on the basis of the results of the comparing step and of the eye gazing parameters of the at least two individuals.