Voice Authentication Thresholds for False Rejection Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Voice authentication systems face challenges in setting appropriate thresholds to balance false rejection and false acceptance rates, as different authentication engines produce varying scores for the same content types, leading to vulnerabilities in security performance.
Innovation Solution
The method involves ascertaining a measure of confidence for each voiceprint through simulated impostor testing, using individual false acceptance and rejection rates to derive an individual equal error rate, and implementing optimization actions such as re-building voiceprints or adjusting thresholds to enhance security performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the threshold score is set high to reduce false acceptance rate, then security is improved, but false rejection rate increases causing service issues
Solution Approach 1:
The patent applies parameter changes by dynamically adjusting the threshold score based on the content type being authenticated. Different content types (account numbers, dates, phrases) have different threshold values assigned to them, allowing the system to optimize security for each type while minimizing false rejections. This resolves the contradiction by making the threshold adaptive rather than fixed.
Solution Approach 2:
The system implements dynamics by making the threshold score flexible and adaptable based on the specific authentication context. The threshold is not static but changes according to the content type being verified, allowing the system to dynamically balance security requirements with service availability for different authentication scenarios.
2Ease of operation
If the threshold score is set low to reduce false rejection rate, then service availability is improved, but false acceptance rate increases compromising security
Solution Approach 1:
The patent resolves this contradiction by changing the threshold parameter based on content type. Each content type has its own optimized threshold that balances security and service availability appropriately, rather than using a single low threshold that would compromise security across all authentication types.
3Measurement precision
If different authentication engines are used to improve measurement capability, then authentication flexibility is improved, but score consistency deteriorates due to varying scores for same content
Solution Approach 1:
The patent addresses score consistency by introducing content-type-specific threshold parameters. Even though different engines produce different scores, the system compensates by adjusting the threshold based on the content type being authenticated, ensuring consistent security outcomes across different engines and content types.
Solution Approach 2:
The system applies local quality by tailoring the threshold parameter to each specific content type. Instead of using a uniform threshold across all authentication scenarios, each content type (account number, date, phrase) receives a locally optimized threshold that accounts for its specific characteristics and the variability introduced by different authentication engines.
4Ease of operation
If a single threshold is used for all content types to simplify operation, then ease of operation is improved, but authentication performance deteriorates due to varying content characteristics
Solution Approach 1:
The patent resolves this contradiction by implementing parameter changes based on content type. The system automatically selects appropriate threshold values for different content types, maintaining ease of operation through automation while achieving high authentication performance through content-specific optimization.
Solution Approach 2:
The system applies self-service by automatically selecting and applying the appropriate threshold for each content type without requiring manual configuration. The authentication engine itself manages the complexity of multiple thresholds, making the system easy to operate while maintaining high performance through content-type-specific parameter selection.
Data Source
AI summary
A method for configuring a voice authentication system comprises ascertaining a measure of confidence associated with a voice sample enrolled with the authentication system. The measure of confidence is derived through simulated impostor testing carried out on the enrolled sample.