Facial Recognition Using Known and Unknown User Image Comparison
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Facial recognition systems face challenges in accurately identifying users without relying on confidence threshold values, which can lead to incorrect identification of unknown users as known users or failure to identify known users due to improper threshold settings.
Innovation Solution
The method compares an image of a user to a collection of images that includes both known and unknown users, determining a higher match with known images over unknown images by calculating confidence values, eliminating the need for a confidence threshold and ensuring accurate user identification.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If a confidence threshold value is used to determine user identification, then the system can operate with a decision criterion, but unknown users may be incorrectly identified as known users or known users may fail to be identified due to improper threshold settings
Solution Approach 1:
The patent removes the confidence threshold value from the identification system. Instead of comparing confidence values against a fixed threshold, the system directly compares the confidence value of the most likely known user against the confidence values of unknown users, eliminating the source of identification errors caused by improper threshold settings
Solution Approach 2:
The patent inverts the traditional approach by not asking 'does the confidence value exceed the threshold?' but rather 'is the known user's confidence value higher than any unknown user's confidence value?'. This reversal of the comparison logic fundamentally changes the identification criterion from threshold-based to relative-ranking-based
2Device complexity
If the collection of images includes only known users, then the system simplifies the identification process, but it cannot prevent incorrect identification of unknown users as known users
Solution Approach 1:
The patent performs preliminary action by obtaining confidence values for both known users and unknown users before making the identification decision. This advance preparation of confidence values for all potential users (known and unknown) enables the system to accurately determine whether a captured image matches a known user or represents an unknown user
Solution Approach 2:
The patent applies different treatment to different parts of the image collection: known users are used for positive identification while unknown users serve as a reference group for negative identification. This local differentiation in function allows the system to maintain simple collection structure while achieving high identification reliability
Data Source
AI summary
A device may receive an image of a user. The device may compare the image to an image of a known user and an image of an unknown user. The device may select, based comparing the image to the image of a known user and the image of an unknown user, one of: the image of the known user, or the image of the unknown user. The device may identify, when the image of the known user is selected, the user as the known user. The device may not identify the user when the image of the unknown user is selected.