Biometric Authentication System Using Voice Models for Secure POS Transactions
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Point-of-sale (POS) transactions using personal identification numbers (PINs) are insecure due to ease of compromise, leading to reduced user trust and verification effectiveness.
Innovation Solution
Implementing biometric authentication systems that utilize unique user biometrics, such as voice recognition, to verify users by comparing received biometric data to stored models within a database, with location-aware techniques to efficiently retrieve and compare biometric data models, reducing the need for extensive database searches.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If PIN-based authentication is used for POS transactions, then the verification process is simple and fast, but security is compromised because PINs are easily stolen or guessed
Solution Approach 1:
The patent replaces the mechanical/PIN-based authentication system with a biometric authentication system using voice recognition. The voice biometric data is captured, processed through speaker verification algorithms, and compared against stored voice models to authenticate users. This substitution eliminates the security vulnerabilities of PINs while maintaining authentication functionality.
Solution Approach 2:
The patent creates and stores copies of users' voice biometric data in the form of voice models within a database. These voice model copies are then used for verification during transactions by comparing newly captured voice samples against the stored models, enabling secure authentication without requiring users to remember PINs.
2Measurement precision
If the entire biometric database is searched for each transaction, then accurate user identification is ensured, but transaction processing time increases significantly
Solution Approach 1:
The patent performs preliminary actions by pre-processing and storing voice biometric data in organized voice models within the database before transactions occur. During transactions, the system can quickly retrieve and compare against these pre-prepared models using references rather than searching the entire raw database, thus maintaining identification accuracy while reducing processing time.
Solution Approach 2:
The patent segments the large biometric database into organized voice model collections with reference structures. Instead of searching through all raw biometric data, the system divides the search space into manageable segments organized by user identifiers, allowing efficient retrieval of specific voice models for comparison during transactions.
Data Source
AI summary
Embodiments of the invention are directed to systems and methods for biometrics transaction processing. A location of a device associated with a user may be determined. A reference to a biometric data model associated with the user stored within a database may be retrieved, based at least in part on the location. Biometric data may be received from the user. Using the reference, the biometric data may be compared to the biometric data model stored within the database. A determination may be made whether the user is authenticated for the transaction based on the comparing step.