Biometric Enrollment for Face Coverings Using Digital Image Warping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Individuals wearing face coverings, such as the Niqāb, are excluded from using facial biometric authentication systems as there is no effective mechanism to enroll biometric information without requiring them to remove their face coverings, hindering identification and verification processes.
Innovation Solution
A method that involves receiving an uncovered face image, generating a fixed-size representation, obtaining a covered face image, and creating a record of both representations, allowing for subsequent biometric identification and verification without requiring the individual to reveal their face, using techniques like head pose estimation and image warping to add artificial face coverings to images.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If facial biometric authentication requires the individual to remove their face covering, then biometric identification accuracy is improved, but accessibility and inclusivity for individuals wearing face coverings deteriorates
Solution Approach 1:
The system performs preliminary action by capturing and storing both covered and uncovered face images during the enrollment phase. This allows the biometric template to be established before the actual authentication occurs, enabling the system to recognize individuals wearing face coverings without requiring them to remove them during authentication.
Solution Approach 2:
The system creates a copy of the face image with an artificial face covering applied to it, stored alongside the original uncovered face image. This copied representation allows the system to match covered face images during authentication against the stored template, enabling accurate identification without requiring the individual to remove their face covering.
2Measurement precision
If the system stores only uncovered face images for biometric enrollment, then biometric data accuracy is improved, but the system cannot verify individuals wearing face coverings
Solution Approach 1:
During the enrollment phase, the system preliminarily captures both the uncovered face image (for accurate biometric data extraction) and the covered face image (for verification during authentication). This preliminary dual-capture allows the system to maintain high accuracy while enabling future verification of individuals wearing face coverings.
Solution Approach 2:
The system uses an artificial face covering as an intermediary element that is digitally applied to the uncovered face image to create a covered face image representation. This intermediary allows the system to store and verify both covered and uncovered representations, bridging the gap between accurate biometric data capture and verification capability.
3Reliability
If the system requires face uncovering for authentication, then biometric verification reliability is improved, but the ease of operation for users wearing face coverings deteriorates
Solution Approach 1:
The system creates a digital copy of the face image with an artificial face covering applied, stored as part of the biometric template. During authentication, the system compares the live covered face image against this stored copy, eliminating the need for the user to remove their face covering while maintaining verification reliability.
Solution Approach 2:
The system changes the parameter of face coverage status by digitally applying artificial face coverings to stored face images, transforming them into covered face representations. This parameter change allows the system to accept and verify both covered and uncovered face images during authentication, improving ease of operation for users wearing face coverings.
Data Source
AI summary
A system for biometric enrollment can include a server including a processor configured to receive an uncovered face image of a subject. The processor can generate a first fixed-size representation (FXR) based on the uncovered face image and a covered face image based on the uncovered face image. The processor can generate a second FXR based on the covered face image. The processor can enroll the subject associated with the uncovered face image by storing the first FXR and the second FXR in a data store.