Facial Recognition Template Storage on Mobile Credential
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Facial recognition systems in access control face challenges with offline enrollment processes, requiring template distribution across networks, which can be insecure and inefficient, especially in ultra-high security deployments where network connectivity is limited, and current biometric data protection regulations restrict wide distribution of facial recognition data.
Innovation Solution
Storing facial templates directly on the user's credential device, allowing secure transmission and comparison with real-time images captured during access attempts, eliminating the need for network distribution and enabling two-factor authentication.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If facial recognition templates are distributed across networks for offline enrollment, then user enrollment can be performed without network connectivity, but security is compromised and data protection regulations are violated
Solution Approach 1:
The patent extracts the facial recognition template storage function from centralized network servers and relocates it to individual credential devices (mobile phones, smart cards). Each user's template is stored locally on their personal device rather than being distributed across network servers, eliminating the security risks of centralized storage while enabling offline operation.
Solution Approach 2:
The patent creates a copy of the facial recognition template on the user's personal credential device. Instead of relying on a single centralized database, the template is replicated on each user's device, allowing them to present their own stored template during authentication without requiring network connectivity to a central server.
2Reliability
If facial recognition templates are stored on credential devices, then security and data protection are improved, but network connectivity is required for template distribution during enrollment
Solution Approach 1:
The patent performs the template distribution and storage setup during the initial enrollment process, which is a preliminary action. Users provide their facial images during enrollment, and the system generates and stores their templates on their personal devices at this time. Once enrolled, users can authenticate without network connectivity, as the template is already stored on their device.
3Measurement precision
If facial recognition systems use deep neural networks for high accuracy, then identification capability is improved, but computational complexity and system resource requirements increase
Solution Approach 1:
The patent extracts the computationally intensive facial recognition processing from the credential device and relocates it to the reader device. The credential device only stores and transmits the template, while the reader device performs the complex deep neural network comparison operations. This division of labor allows high accuracy without requiring complex computational resources on the user's personal device.
Solution Approach 2:
The patent introduces the reader device as an intermediary between the credential device and the authentication decision. The reader device receives the template from the credential device, performs the complex facial recognition comparison using deep neural networks, and makes the authentication decision. This intermediary handles the computational complexity while keeping the credential device simple.
Data Source
AI summary
A non-transitory computer readable medium comprising executable code that causes one or more processors to: receive, at a reader device from a credential device of a user, at least one of a first image of the user or a first representation of a face of the user; if a first image of the user was received, then generate, at the reader device, a generated representation of the face of the user using the first image; capture, at the reader device, a second image of the user and generate a second representation of the face of the user using the second image; determine a likelihood of a match between the second representation and at least one of the first representation and the generated representation; and permit access by the user to a secure asset in instances where the likelihood of a match meets a threshold.