Line-of-Sight Detection Calibration with Selective Data Registration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for detecting a line of sight suffer from reduced accuracy due to individual differences in eyeball shape, leading to deviations in detected directions, which can be compensated by calibrating with calibration data but may reduce accuracy if the data accuracy is low.
Innovation Solution
An information processing apparatus with a line-of-sight estimator, correction amount calculator, and registration determination section that calculates and registers calibration data based on parameters related to objects within a specified angular range, ensuring high accuracy by selecting reliable data for calibration.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If calibration data is collected using existing methods, then calibration can be performed to compensate for detection deviations, but the accuracy in detecting line of sight is reduced when the data accuracy is low
Solution Approach 1:
The system performs preliminary evaluation of calibration data quality before final registration. The registration determination section assesses parameters such as the number of objects within angular range, angular spacing, object distances, and noticeability scores before committing calibration data to storage, ensuring only high-quality data is registered
Solution Approach 2:
The system establishes a feedback loop where calibration data is evaluated against multiple parameters, and only data meeting predetermined thresholds is registered. This feedback mechanism prevents low-quality data from degrading the overall calibration accuracy
2Quantity of substance
If multiple objects are considered within the angular range, then more calibration data can be collected, but the reliability of calibration data decreases due to potential selection of inappropriate objects
Solution Approach 1:
The system changes multiple parameters simultaneously to evaluate data quality: it considers the number of objects within angular range, angular spacing between objects, distances from point of view to objects, and noticeability scores. By adjusting and evaluating against multiple parameters, the system can process more objects while maintaining reliability through comprehensive quality assessment
3Productivity
If calibration data is registered without strict selection criteria, then the calibration process can be completed faster, but the accuracy in detecting line of sight is reduced
Solution Approach 1:
The system performs preliminary quality assessment of calibration data before registration by evaluating multiple parameters including object count, angular spacing, distances, and noticeability. This preliminary action ensures that accuracy requirements are met before the calibration process completes, preventing low-accuracy data from being registered
Solution Approach 2:
The registration determination section autonomously evaluates calibration data quality against predetermined criteria and makes self-service decisions about registration. This automated quality control maintains accuracy standards without requiring manual intervention, balancing speed and precision
Data Source
AI summary
Provided is an information processing apparatus which includes a line-of-sight estimator, a correction amount calculator, and a registration determination section. The line-of-sight estimator calculates an estimation vector obtained by estimating a direction of a line of sight of a user. The correction amount calculator calculates a correction amount related to the estimation vector on the basis of at least one object that is within a specified angular range that is set using the estimation vector as a reference. The registration determination section determines whether to register, in a data store, calibration data in which the estimation vector and the correction amount are associated with each other, on the basis of a parameter related to the at least one object within the specified angular range.