Biometric Identity Verification Using Face Search and Personal Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current identity verification methods for electronic wallet services, relying on phone numbers or email addresses, are inaccurate as they cannot uniquely identify individuals, leading to financial risks due to the ability of users to register multiple accounts under different names.
Innovation Solution
A device and method utilizing biometric data, specifically face images, in conjunction with personal information, to conduct a face search in a historical user database, generating a similarity score to accurately verify users and prevent fraudulent account registrations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If phone number or email address based identity verification is used, then the verification process is simple and easy to operate, but the verification accuracy is low and cannot uniquely identify individuals
Solution Approach 1:
The patent combines biometric data (face images) with personal information (phone numbers, email addresses, names) to create a comprehensive identity verification system. This merging of multiple identification methods allows the system to maintain ease of operation through familiar personal information while adding biometric verification to significantly improve accuracy and prevent fraudulent account registrations
Solution Approach 2:
The patent introduces a biometric intermediary (face recognition technology) that acts as a mediator between the user and the identity verification system. The face image serves as an objective, unique identifier that bridges the gap between simple personal information and accurate identity confirmation, enabling the system to verify identities without requiring complex user actions
2Measurement precision
If biometric data and face search are implemented, then identity verification accuracy is improved, but device complexity increases
Solution Approach 1:
The patent implements a universal identity verification system that handles both new user registrations and historical user verifications through a single integrated platform. The system maintains a historical user database that stores biometric and personal information, allowing the same verification mechanism to serve multiple functions: identifying new users, verifying returning users, and detecting fraudulent registrations across different contexts
Solution Approach 2:
The patent performs preliminary actions by pre-processing and storing biometric data (face images) and personal information in a historical user database during initial user registrations. This preliminary data collection and organization enables rapid, accurate verification during subsequent registrations without requiring complex real-time processing, thus reducing operational complexity while maintaining high verification accuracy
3Reliability
If face search and similarity scoring are conducted, then fraudulent account registrations are prevented, but processing time increases
Solution Approach 1:
The patent applies partial action by implementing a tiered verification approach: the system first performs a quick face search to identify potential matches, then conducts detailed similarity scoring only for candidates that meet initial thresholds. This partial application of comprehensive verification maintains high fraud prevention capability while avoiding the time cost of exhaustive analysis for every registration attempt
Data Source
AI summary
A computer-implemented method includes obtaining user data associated with a first user, the user data including biometric data associated with the first user and personal information of the first user; conducting a biometric similarity search in a historical user database based on the biometric data; determining, based on the biometric similarity search, that multiple historical users have similar biometric data to the first user; retrieving corresponding personal information of each of the multiple historical users; generating a respective personal information similarity score for each of the multiple historical users; and identifying, as the first user, a historical user of the multiple historical users, the personal information similarity score corresponding to the identified historical user indicating a higher similarity than the personal information similarity scores corresponding to other historical users of the multiple historical users.