Face Enrolment Verification Using Swapped-Position Scene Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing face recognition systems for enrolling unauthorized persons in a database of authorized individuals are inefficient and prone to errors due to manual intervention and misinterpretation of single-image analysis, especially when customers appear younger than their true age.
Innovation Solution
A database enrolment system that requires an authorized person to swap places with an unauthorized person in the camera's field of view, using face recognition to detect and confirm the faces in two subsequent scenes, reducing the likelihood of erroneous entries by ensuring a specific sequence of events is detected.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If manual enrolment procedure is used to photograph and add unauthorised person information to the database, then the enrolment can be performed, but the procedure is lengthy and inefficient
Solution Approach 1:
The patent replaces the manual mechanical enrolment process (physically photographing and entering data) with an automated optical system. The camera captures images and the processing unit automatically extracts facial features and enrols the person in the database, eliminating the need for manual data entry and significantly reducing enrolment time while maintaining accuracy.
2Ease of operation
If single image analysis is used to detect enrolment request, then the system is simple to operate, but the likelihood of misinterpreting the request and erroneous entries is high
Solution Approach 1:
The patent segments the enrolment verification process into multiple distinct stages: detecting the initial scene with both persons, verifying their identities against the database, detecting the swapped positions scene, and confirming the enrolment request. This multi-stage segmentation allows the system to maintain operational simplicity while dramatically improving reliability by cross-validating enrolment requests through sequential verification steps.
Solution Approach 2:
The system incorporates feedback mechanisms by continuously monitoring the camera feed for the swapped positions scene after detecting the initial enrolment request. The processing unit provides feedback by verifying identities, checking for database matches, and confirming the sequence of events before finalising enrolment, thereby reducing erroneous entries while maintaining ease of operation.
3Productivity
If authorised person manually verifies customer age on subsequent visits, then age-restricted products can be protected, but the process is inconvenient and inefficient
Solution Approach 1:
The patent applies preliminary action by enrolling customers in the database during their first verified purchase. Once enrolled, their facial biometric data is stored and automatically recognised on subsequent visits, eliminating the need for repeated manual age verification. This preliminary enrolment action streamlines future transactions, improving both verification efficiency and customer convenience while maintaining product protection.
Data Source
Figure 1a~1b
AI summary
A database enrolment system for enrolling an unauthorised person in a database of authorised persons is provided. The system comprises: a camera for imaging a field of view, and a processing unit in communication with the database of authorised persons. The database contains information relating to the faces of authorised persons. The processing unit comprises a face recognition subsystem and is configured to perform an enrolment routine for enrolling an unauthorised person in the database. The enrolment routine comprises: detecting an initialisation scene, detecting a subsequent confirmation scene, and when the initialisation scene and the confirmation scene are detected, transmitting enrolment information relating to the face of the unauthorised person to the database for enrolment as an authorised person. Detecting initialisation scene includes: using the face recognition subsystem to detect the face of an authorised person in the field of view and to detect the face of an unauthorised person in the field of view, wherein the face of the authorised person is detected on a first side of the face of the unauthorised person in the field of view. Detecting the subsequent confirmation scene includes: using the face recognition subsystem to detect the face of the authorised person and the face of the unauthorised person in the field of view, wherein the face of the authorised person is detected on a second side of the face of the unauthorised person in the field of view.