Adaptive Illumination for Face Recognition Reliability
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Face recognition technology in interactive terminals suffers from low reliability due to unpredictable lighting conditions and user positioning, making it difficult to enforce control-based solutions like lighting or pose/expression control, which can be unfavorable for users.
Innovation Solution
A method that adjusts illumination levels based on the reliability measure of the face recognition algorithm, incrementally increasing lighting only when necessary to achieve reliable identification, and personalizes the terminal's screen upon successful identification, using a feedback loop to optimize lighting levels without dazzling the user.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If control-based solutions (lighting or pose/expression control) are enforced on users in interactive terminals, then face recognition reliability is improved, but user experience deteriorates due to unfavorable conditions
Solution Approach 1:
The system performs an initial face recognition attempt, evaluates the reliability of the result, and uses this feedback to determine whether to adjust lighting. This closed-loop feedback mechanism allows the system to adaptively improve recognition reliability only when necessary, rather than enforcing control measures on all users regardless of actual need.
Solution Approach 2:
The system applies lighting adjustment only partially - specifically, only when the initial face recognition reliability is below a predetermined threshold. This partial action approach avoids unnecessary lighting adjustments for users who already meet the reliability threshold, thereby preserving user experience while still improving reliability when needed.
2Reliability
If lighting level is increased to improve face recognition reliability, then identification accuracy is improved, but user comfort deteriorates due to potential dazzling
Solution Approach 1:
The system applies lighting adjustment only partially - specifically, only when the initial face recognition reliability is below a predetermined threshold. This ensures that lighting is increased only to the extent necessary to achieve reliable identification, avoiding excessive lighting that would cause user discomfort or dazzling.
Solution Approach 2:
The system uses feedback from the initial recognition attempt to determine whether lighting adjustment is needed. This feedback-driven approach ensures that lighting is adjusted only when the reliability metric indicates insufficient identification accuracy, thereby avoiding unnecessary lighting increases that would harm user comfort.
3Reliability
If adaptive lighting adjustment is implemented based on reliability measures, then face recognition reliability is improved, but device complexity increases
Solution Approach 1:
The system implements adaptive lighting adjustment in a partial manner - only when reliability thresholds are not met. This selective implementation reduces the overall complexity burden compared to a system that continuously adjusts lighting based on multiple factors, while still achieving the goal of improved recognition reliability.
Solution Approach 2:
The system uses a straightforward feedback mechanism where the reliability measure from initial recognition directly controls the lighting adjustment decision. This simple feedback loop avoids the need for complex multi-factor decision algorithms, thereby limiting the increase in device complexity while still achieving adaptive lighting adjustment.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
This invention provides a method of face recognition and a device for implementing said method, the method comprising the steps of: (a) performing a face recognition algorithm on an image of a user to determine an identity of the user, the image having an illumination level; (b) determining if a reliability measurement of the identity of the user determined by the face recognition algorithm meets a threshold indicating reliable identification and, if the reliability measurement is below the threshold indicating reliable identification; (c) receiving a new image of the user, the new image having an adjusted illumination level; and (d) performing a further face recognition algorithm on the new image of the user to determine the identity of the user.