Biometric Threshold Automation via Cumulative Histograms
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Biometric systems face challenges in accurately setting a match threshold to achieve desired False Match Rate (FMR) and False Non-Match Rate (FNMR) levels, as the current process is manual and imprecise, lacking automation and data consideration to optimize error rates.
Innovation Solution
The system generates databases of match scores using imposter and genuine sample pairs, creating cumulative histogram data tables to determine false match and non-match rates, allowing users to select and apply thresholds for desired FMR and FNMR levels, and incorporates false match probability scores to enhance decision-making.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a manual threshold setting process is used, then the system is simple to operate, but the precision of FMR and FNMR control is insufficient
Solution Approach 1:
The system performs preliminary actions by generating databases of match scores from imposter and genuine sample pairs before threshold setting. Cumulative histogram data tables are created in advance, containing pre-calculated FMR and FNMR values for various threshold levels. This preliminary preparation enables precise threshold selection without complex real-time calculations, resolving the contradiction between precision and complexity.
Solution Approach 2:
The system serves itself by automatically generating match score databases and cumulative histogram tables using its own biometric data. The false match probability scores are self-calculated based on the distribution of match scores from the system's genuine and imposter samples. This self-service approach eliminates the need for external manual calibration while maintaining precision.
2Reliability
If automated threshold setting with comprehensive data analysis is implemented, then the FMR and FNMR control is optimized, but the processing time and computational resources increase
Solution Approach 1:
Cumulative histogram data tables are generated in advance during system initialization or offline processing. These tables contain pre-computed FMR and FNMR values for multiple threshold levels, derived from databases of match scores between imposter and genuine sample pairs. When threshold setting is needed, the system simply queries these pre-computed tables rather than performing complex real-time analysis, significantly reducing time loss while maintaining high reliability.
Solution Approach 2:
The system uses a representative subset of sample pairs to generate match score databases rather than exhaustively analyzing all possible combinations. By selecting sufficient genuine and imposter samples to capture the distribution characteristics, the system achieves reliable FMR and FNMR estimation without the excessive computational burden of complete enumeration, balancing reliability and time efficiency.
3Measurement precision
If false match probability scores are incorporated into the decision process, then the confidence in match decisions is improved, but the device complexity increases
Solution Approach 1:
False match probability scores serve as an intermediary metric between the raw match score and the final match/no-match decision. Instead of directly comparing match scores to a fixed threshold, the system uses the cumulative histogram data to determine the probability that a given match score represents a false match. This intermediary probability layer provides more nuanced confidence information while adding only a simple lookup operation to the decision process, minimizing complexity increase.
Solution Approach 2:
The system incorporates feedback by using the distribution of match scores from genuine and imposter samples to dynamically adjust the interpretation of individual match scores. The cumulative histogram data provides feedback on how unusual a given match score is, allowing the system to assign appropriate confidence levels. This feedback mechanism improves match confidence accuracy without requiring complex adaptive algorithms, as it simply compares each match score against the pre-established distribution patterns.
Data Source
AI summary
In real biometric systems, false match rates and false non-match rates of 0% do not exist. There is always some probability that a purported match is false, and that a genuine match is not identified. The performance of biometric systems is often expressed in part in terms of their false match rate and false non-match rate, with the equal error rate being when the two are equal. There is a tradeoff between the FMR and FNMR in biometric systems which can be adjusted by changing a matching threshold. This matching threshold can be automatically, dynamically and/or user adjusted so that a biometric system of interest can achieve a desired FMR and FNMR.