Eye Tracking Alternating Resolution and Illumination Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current eye/gaze tracking systems face challenges in maintaining robustness and stability due to issues with image resolution, light source positioning, and hardware-related problems, particularly when dealing with reflections and varying lighting conditions.
Innovation Solution
The system employs a control unit to manage a sequence of image frames with alternating resolutions and coordinated illumination, using both full-frame and region-of-interest images to enhance feature recognition and gaze tracking accuracy, while suppressing unfavorable reflections and ensuring testability and repeatability.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If image cropping/ROI is applied to decrease bandwidth and increase read-out frequency, then the effects of non-controlled light sources are reduced, but the image size becomes small and may cause the tracker to lock onto false eye candidates
Solution Approach 1:
The system alternates between capturing full-field images and ROI images in a periodic sequence. Full-field images are captured periodically to verify eye position and prevent false locking, while ROI images are captured more frequently for high-precision tracking. This periodic switching resolves the contradiction by ensuring both comprehensive coverage and detailed measurement.
Solution Approach 2:
The system dynamically adjusts the field of view by switching between full-frame and cropped ROI modes based on tracking confidence and eye position stability. When tracking is stable, the system uses ROI for precision; when uncertainty arises, it expands to full-field view to verify eye location, thus adapting the image area to current needs.
2Reliability
If multiple light sources are used to improve gaze tracking, then unfavorable reflections can be suppressed, but hardware-related stability problems and visible flicker may occur
Solution Approach 1:
Multiple light sources are activated in a periodic, alternating sequence rather than simultaneously or randomly. Each light source is turned on for a specific duration and then switched off, creating a time-multiplexed illumination pattern. This periodic activation suppresses reflections from any single light source while avoiding the stability issues and flicker that would result from continuous or random switching.
Solution Approach 2:
The system pre-plans the activation sequence of multiple light sources to ensure optimal illumination while avoiding harmful reflections. By determining in advance which light source should be active at each moment, the system prevents instability and flicker before they occur, rather than reacting to problems as they arise.
3Adaptability or versatility
If feature recognition algorithms are used to identify objects, then the camera's full field of view is beneficial, but switching between wide field of view and ROI increases system complexity
Solution Approach 1:
The system employs a periodic sequence that alternates between full-field imaging for feature recognition and ROI imaging for precise eye tracking. This regular switching pattern simplifies control logic compared to dynamic, condition-based switching, as the system follows a predetermined temporal pattern rather than requiring complex real-time decision-making.
Data Source
AI summary
At least one image registering unit records at least one series of images representing a subject. A control unit controls an operation sequence for the at least one image registering unit in such a manner that a subsequent data processing unit receives a repeating sequence of image frames there from, wherein each period contains at least one image frame of a first resolution and at least one image frame of a second resolution being different from the first resolution. Based on the registered image frames, the data processing unit produces eye/gaze tracking data with respect to the subject.