Biometric Authentication Probability Fusion for Spoof-Resistant Login
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Biometric authentication systems are vulnerable to spoofing attacks where counterfeit biometric information is used to deceive the system, leading to incorrect authentication.
Innovation Solution
An information processing device that calculates a probability of genuine identity based on a spoofing score and collation score, using methods like deep learning and machine learning to determine the likelihood of spoofing, thereby improving authentication accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If biometric authentication is performed using only collation score, then authentication convenience is improved, but authentication reliability deteriorates due to spoofing attacks
Solution Approach 1:
The authentication system is segmented into multiple independent scoring components: collation score calculation unit, spoofing score calculation unit, and probability calculation unit. Each unit processes different aspects of biometric verification independently, allowing the system to maintain convenience while improving reliability through diversified evaluation criteria that can detect spoofing attempts.
2Reliability
If spoofing detection is added to biometric authentication, then authentication reliability is improved, but device complexity increases
Solution Approach 1:
The spoofing detection functionality is merged with the existing collation score calculation into a unified authentication processing unit. The spoofing score calculation unit shares computational resources and processing pipelines with the collation unit, allowing both functions to operate simultaneously without requiring completely separate hardware systems, thus limiting the increase in device complexity.
Solution Approach 2:
The authentication processing unit is designed with multi-functionality to perform both collation scoring and spoofing detection using the same biometric input data. This universal approach allows a single system component to serve multiple authentication purposes, reducing the need for additional dedicated hardware and minimizing overall system complexity.
3Measurement precision
If probability calculation based on multiple scores is performed, then measurement precision is improved, but calculation time increases
Solution Approach 1:
The system performs preliminary calculations by pre-establishing the relationship between collation scores, spoofing scores, and authentication probabilities through probability calculation unit training during system initialization. This preliminary action allows the system to use pre-computed probability mappings during actual authentication, reducing real-time calculation time while maintaining high measurement precision.
Solution Approach 2:
The probability calculation unit transforms the complex multi-dimensional scoring problem into a simplified parameter relationship by calculating authentication probability as a function of the collation and spoofing scores. This parameter transformation approach converts multiple score evaluations into a single probability value, improving measurement precision while reducing the computational burden and time required for final authentication decisions.
Data Source
AI summary
An information processing device calculates a probability that biometric information corresponds to a subject, or the biometric information does not correspond to the subject, based on a first score indicating a degree to which the subject is spoofed and a second score indicating the degree of similarity between the biometric information and registered biometric information, and performs authentication related to the subject based on the probability.