Eye Tracking Glint Identification via Optical Pattern Matching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Eye tracking systems face challenges in accurately determining gaze direction when users wear glasses under VR glasses due to reflections from optical arrangements, which can lead to errors in identifying pupil and glint positions.
Innovation Solution
The method involves identifying glints in images captured by an image sensor, determining their relative positions, and distinguishing between reflections from the eye's cornea and optical arrangements based on known patterns, such as Fresnel lens rings, to accurately determine gaze direction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If illuminators are used to illuminate the user's eyes for eye tracking, then the gaze direction can be determined based on pupil and glint identification, but reflections from optical arrangements (such as glasses worn under VR glasses) may be mistaken for corneal reflections, introducing errors in eye tracking
Solution Approach 1:
The patent segments the identification process into multiple stages: first identifying all potential glints, then classifying them as either corneal reflections or optical arrangement reflections based on pattern matching. This segmentation allows the system to handle different types of reflections separately, improving measurement precision by correctly identifying and excluding false glints from optical arrangements.
Solution Approach 2:
The patent performs preliminary identification of all glints before determining which ones are valid for gaze calculation. By预先 identifying and classifying reflections from optical arrangements based on their known patterns, the system prevents these false glints from interfering with subsequent gaze direction calculations, thereby improving measurement accuracy.
2Loss of information
If the image sensor captures reflections from both the eye's cornea and optical arrangements, then more glints are available for analysis, but the presence of optical arrangement reflections complicates the identification of valid corneal reflections for eye tracking
Solution Approach 1:
The patent uses the known pattern of optical arrangement features as a template or copy to identify and distinguish their reflections in the captured image. By comparing detected glints against the expected pattern from optical arrangements, the system can automatically identify and exclude these false reflections, maintaining information availability while managing identification complexity through pattern recognition.
3Adaptability or versatility
If users wear glasses with optical components under VR glasses, then the eye tracking system must handle additional reflections, but this reduces the reliability of gaze direction determination
Solution Approach 1:
The patent converts the harmful effect of optical arrangement reflections into a beneficial identification mechanism. By using the known pattern of optical features as a signature, the system transforms the presence of these reflections from a source of error into a detectable pattern that helps identify and exclude false glints, thereby maintaining eye tracking reliability even when users wear glasses.
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 reduces errors in eye tracking by correctly identifying reflections from optical arrangements, enhancing the accuracy of determining gaze direction, especially in scenarios where users wear glasses with optical components like Fresnel lenses.
Implementation Method 1
The image is a result of the image sensor detecting light from one or more illuminators reflected from the eye of the user and reflected from optic arrangements located between the one or more illuminators and the eye of the user
Data Source
Figure 1
Figure 2~3a
Figure 3b~3c
AI summary
A method of identifying reflections from optical arrangements in an eye tracking system and a corresponding eye tracking system is disclosed. An image of an eye of a user is received from an image sensor, the image resulting from the image sensor detecting light from one or more illuminators reflected from the eye of the user and reflected from optic arrangements located between the one or more illuminators and the eye of the user, wherein the optic arrangements comprise features having a known pattern. Two or more glints are identified in the image, wherein a glint is a representation in the image of a reflection of light from a cornea of the eye of the user and/or from a feature of the optic arrangements. Relative positions in the image of the identified glints are determined, and on condition that the relative positions in the image of the identified glints conform with the known pattern of the features of the optic arrangements, at least one glint of the identified glints is identified as being a representation in the image of a reflection from a feature of the optic arrangements.