Eye Gaze Detection Using Corneal Reflection Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The corneal reflection method for eye gaze estimation is hindered by outliers such as unexpected light sources and reflections from contact lenses, which can lead to inaccurate identification of corneal reflection images, affecting the accuracy of eye gaze vector estimation.
Innovation Solution
An information processing apparatus and method that estimates the position of the eyeball and cornea from time-series images, identifies candidate corneal reflection images, and detects these images using a detection unit, thereby improving the accuracy of eye gaze vector estimation by excluding outliers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the corneal reflection method is used to estimate eye gaze, then the eye gaze can be measured using corneal reflection images, but outliers such as unexpected light sources and contact lens reflections may be falsely identified as corneal reflection images, reducing measurement accuracy
Solution Approach 1:
The system performs preliminary actions by estimating the eyeball center position and cornea center position before detecting corneal reflection images. These preliminary estimations establish expected spatial relationships and constraints that are used to validate potential reflection candidates, preventing outliers from being falsely identified
Solution Approach 2:
The system uses feedback by comparing the actual positions of detected light sources and potential reflection images against the estimated geometric relationships ( eyeball center, cornea center, and reflection point relationships). This feedback mechanism allows the system to distinguish true corneal reflections from outliers based on whether they conform to the expected geometric constraints
2Measurement precision
If multiple light sources are used to illuminate the eye, then more corneal reflection images can be obtained for better estimation, but the risk of including outlier reflections from unexpected light sources increases
Solution Approach 1:
The system establishes expected spatial relationships between multiple light sources, eyeball center, cornea center, and their corresponding reflection points. By using feedback from these geometric constraints, the system can identify and exclude reflections that do not conform to the expected relationships, even when multiple light sources are present
Solution Approach 2:
The estimated eyeball center and cornea center positions act as intermediaries that mediate between the light sources and the reflection detection process. These intermediaries provide a geometric framework that helps distinguish true corneal reflections from outliers by verifying whether detected reflections are consistent with the established geometric relationships
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach enhances the detection accuracy of corneal reflection images and improves the precision of eye gaze vector estimation by effectively filtering out outliers and associating light sources with their corresponding reflections.
Implementation Method 1
detect a corneal reflection image corresponding to light from a light source reflected at a cornea from a captured image
Data Source
AI summary
Provided is an information processing apparatus including: a detection unit configured to detect a corneal reflection image corresponding to light from a light source reflected at a cornea from a captured image in which an eye irradiated with the light from the light source is imaged. The detection unit estimates a position of a center of an eyeball on the basis of a plurality of time-series captured images each of which is the captured image according to the above, estimates a position of a center of the cornea on the basis of the estimated position of the center of the eyeball, estimates a position of a candidate for the corneal reflection image on the basis of the estimated position of the center of the cornea, and detects the corneal reflection image from the captured image on the basis of the estimated position of the candidate for the corneal reflection image.