Adaptive Eye Tracker Using Dual Camera Modes
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional eye trackers face challenges in accuracy, speed, reliability, and cost efficiency due to limitations in image quality, obscuration by objects, and the need for simultaneous bright-pupil and dark-pupil imaging modes, which are often compromised by time delays and reduced accuracy when using multiple reference illuminators.
Innovation Solution
An adaptive eye tracker system with at least one coaxial and one non-coaxial camera and illuminator configuration, where a controller selects the best camera based on image quality metrics, allowing for simultaneous imaging in different modes and reducing obscuration risks, while using point-shaped illuminators for improved contrast and dual-camera mode for enhanced accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If multiple reference illuminators are used simultaneously for bright-pupil and dark-pupil imaging modes, then both imaging modes can be supported, but coexisting glints mutually blur the measurements and reduce accuracy
Solution Approach 1:
The system uses time-interlaced illumination where the first and second reference illuminators are activated alternately in periodic cycles. During each cycle, one illuminator is active while the other is inactive, allowing sequential capture of bright-pupil and dark-pupil images without simultaneous glint interference. This periodic switching resolves the contradiction by enabling multi-mode adaptability while maintaining measurement precision through temporal separation.
2Measurement precision
If reference illuminators are used alternately by time interlacing, then simultaneous glint interference is avoided, but a small time delay separates the two images reducing accuracy particularly during saccades
Solution Approach 1:
The system maintains continuous useful action by operating the cameras at very high frame rates (e.g., 1000 fps or higher) during time-interlaced illumination. This ensures that even though illumination switches alternately, the temporal gap between capturing bright-pupil and dark-pupil images is minimized to microsecond levels, effectively maintaining image continuity and accuracy during rapid eye movements like saccades.
3Measurement precision
If a high-performance camera is used to achieve sufficient resolution for accurate eye tracking, then image quality improves, but the cost of the product increases
Solution Approach 1:
The system segments the imaging function by using two separate cameras instead of one high-performance camera. Each camera can be optimized for specific imaging conditions (one for bright-pupil mode, one for dark-pupil mode), allowing the use of simpler, lower-cost cameras that each handle a specific task, thereby reducing overall system cost while maintaining the required measurement precision for eye tracking.
4Device complexity
If the camera is positioned coaxially with the reference illuminator for bright-pupil imaging, then the optical path is simplified, but the line of sight may be obscured by objects such as spectacle frames, eyelashes, and eyebrows
Solution Approach 1:
The system employs asymmetric positioning where the first camera is coaxial with the first reference illuminator for bright-pupil imaging, while the second camera is positioned at a non-coaxial angle relative to the second reference illuminator for dark-pupil imaging. This asymmetric arrangement allows the second camera to capture images from an oblique angle that avoids obscuration by facial features, thereby improving line of sight reliability while maintaining optical path simplicity through dedicated coaxial alignment for each camera-illuminator pair.
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
The system enhances eye tracking accuracy and reliability by ensuring continuous operation even when one line of sight is obscured, reduces energy consumption, and maintains high image quality with simpler cameras, allowing for more flexible camera placement and improved pupil center location determination.
Implementation Method 1
the virtual image of the reference illuminator formed by reflection in the cornea is shrunk by a factor 100 or more
Implementation Method 2
a retinal retro-reflection complements the iris image
Implementation Method 3
a corneo-scleral reflection complements the iris image
Data Source
AI summary
A method includes obtaining an image of an eye in a bright-pupil imaging mode in which a retinal retro-reflection complements the image of the eye in the bright-pupil mode, and obtaining an image of the eye in a dark-pupil imaging mode in which a corneo-scleral retro-reflection complements the image of the eye in the dark-pupil mode. One of the bright-pupil imaging mode and the dark-pupil imaging mode is selected based on quality of obtained images.