Gaze Tracking Scene Camera Reference Location Identification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing gaze tracking systems face challenges in accurately tracking eye movements due to the need for high spatial resolution and the loss of tracking ability with wide-ranging head movements, especially when cameras and light sources are positioned at a distance from the eyes.
Innovation Solution
The system employs an unobtrusive scene camera mounted on eyewear or headwear to identify reference locations within the environment using image recognition techniques, combining scene reference locations and eye-tracking data to determine the objects being viewed, with optional use of intentionally placed reflective surfaces and controlled illumination to enhance tracking accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If cameras and light sources are positioned at a distance from the eyes, then the system becomes less obtrusive, but measurement precision deteriorates due to the need for higher spatial resolution and loss of tracking ability with wide head movements
Solution Approach 1:
The patent uses a scene camera to capture images of the environment and creates a digital representation (copy) of the scene with reference locations. This allows the system to track gaze by comparing eye position against this digital scene copy, maintaining accuracy without requiring the camera to be positioned close to the eye, thus reducing obtrusiveness while preserving measurement precision.
2Measurement precision
If cameras are positioned close to the eyes, then measurement precision improves, but the system becomes more obtrusive
Solution Approach 1:
The patent introduces a scene camera and reference location system as an intermediary. Instead of requiring the eye-tracking camera to be positioned close to the eye, the system uses the scene camera to capture environmental references and the eye-tracking camera to capture eye position, then computationally combines these to achieve accurate gaze tracking with the camera positioned at a comfortable distance from the eye.
3Adaptability or versatility
If wide-ranging head movements are allowed, then adaptability improves, but measurement precision deteriorates due to loss of tracking ability
Solution Approach 1:
The patent implements a feedback mechanism where the scene camera continuously captures images of reference locations in the environment, and the system computationally determines the device's position and orientation relative to these references. This feedback loop allows the system to maintain accurate gaze tracking even during wide head movements by continuously updating the spatial relationship between the device and the environment.
4Measurement precision
If higher spatial resolution is required for accurate eye tracking, then measurement precision improves, but device complexity increases
Solution Approach 1:
The patent applies partial action by using a scene camera to capture only the necessary reference location information from the environment, rather than requiring the entire scene to be captured at high resolution. The eye-tracking camera captures only the eye region, and the computational system processes only the relevant portions of these images to determine gaze position, reducing overall device complexity while maintaining measurement precision.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A system is provided for identifying reference locations within the environment of a device wearer. The system includes a scene camera mounted on eyewear or headwear coupled to a processing unit. The system may recognize objects with known geometries that occur naturally within the wearer's environment or objects that have been intentionally placed at known locations within the wearer's environment. One or more light sources may be mounted on the headwear that illuminate reflective surfaces at selected times and wavelengths to help identify scene reference locations and glints projected from known locations onto the surface of the eye. The processing unit may control light sources to adjust illumination levels in order to help identify reference locations within the environment and corresponding glints on the surface of the eye. Objects may be identified substantially continuously within video images from scene cameras to provide a continuous data stream of reference locations.