Adaptive Gaze Detection Using Dynamic Facial Reference Points
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Eye gaze tracking technologies face challenges in achieving accuracy under varying lighting conditions and head movements, particularly in mobile device and retail settings, due to issues like glints and reflections.
Innovation Solution
A method that detects spatial information from facial reference points such as eye center localization, nostril position, and head pose, assigning weights based on head position and lighting conditions to generate adaptive gaze detection parameters, and dynamically shifting these points to improve accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional gaze detection methods are used, then the system is simple to implement, but accuracy deteriorates under varying lighting conditions and head movements
Solution Approach 1:
The patent applies dynamics by making the reference point selection adaptive and dynamic rather than fixed. The system dynamically adjusts which facial reference points (eye corners, nostrils, lips) are used based on detected head movement and lighting conditions. This allows the gaze detection system to maintain high accuracy across varying conditions without requiring a completely complex reconfiguration of the entire detection architecture.
Solution Approach 2:
The system changes parameters by adjusting the weight and selection of different facial reference points based on head pose and lighting conditions. Instead of using a fixed set of reference points, the system modifies which points are prioritized (e.g., emphasizing nostril position when head moves, or eye corners under certain lighting) to maintain measurement precision while managing system complexity.
2Measurement precision
If multiple facial reference points are used to improve accuracy, then gaze detection precision improves, but the complexity of processing and weighting multiple points increases
Solution Approach 1:
The patent applies local quality by assigning different weights and processing priorities to different facial reference points based on their local characteristics and the current operational conditions. For example, eye corners may be given higher weight under certain lighting conditions while nostrils are emphasized during head movements. This localized optimization improves overall accuracy without requiring equally complex processing of all points at all times.
Solution Approach 2:
The system uses partial action by selectively activating and weighting only the necessary subset of reference points for each specific situation rather than always processing all possible facial points with equal complexity. This reduces the average processing complexity while maintaining accuracy when needed.
3Reliability
If the system adapts to varying lighting conditions and head movements, then measurement reliability improves, but the computational requirements and processing time increase
Solution Approach 1:
The patent applies preliminary action by pre-establishing the framework for adaptive reference point selection and the relationships between different facial points. The system prepares the weighting mechanisms and detection algorithms in advance so that when head movement or lighting changes are detected, the system can quickly adjust which reference points to use without performing complex real-time calculations for every parameter adjustment.
Solution Approach 2:
The system uses feedback by continuously monitoring head pose and lighting conditions, then adjusting the selection and weighting of facial reference points accordingly. This feedback loop improves reliability by adapting to changing conditions, while the efficiency comes from using the feedback to make targeted adjustments rather than reprocessing all detection parameters from scratch.
4Measurement precision
If dynamic weighting of reference points is implemented, then accuracy under varying conditions improves, but the complexity of weight assignment and parameter generation increases
Solution Approach 1:
The patent applies parameter changes by dynamically adjusting the weights of different facial reference points based on detected head pose and lighting conditions. The system changes which parameters (reference point weights) are emphasized in different situations, allowing accurate gaze detection across varying conditions without requiring a fundamentally more complex weight assignment mechanism.
Solution Approach 2:
The system uses dynamics by making the reference point weighting adaptive rather than static. The weights are adjusted in real-time based on head movement and lighting changes, but within a predefined framework that limits the complexity of the adjustment mechanism. This dynamic approach improves accuracy while managing complexity through structured adaptability.
Data Source
AI summary
Methods, systems, and computer program products for gaze point detection using dynamic facial reference points under varying lighting conditions are provided herein. A computer-implemented method includes detecting items of spatial information pertaining to at least one eye of an individual gazing at one or more objects; detecting an item of spatial information pertaining to at least one nostril of the individual; detecting items of spatial information pertaining to the head of the individual; assigning a distinct weight to each of (i) the spatial information pertaining to the at least one eye and (ii) the spatial information pertaining to the at least one nostril based on the spatial information pertaining to the head; and generating gaze detection parameters applicable to the individual based on (i) the weighted spatial information pertaining to the at least one eye and (ii) the weighted spatial information pertaining to the at least one nostril.