Biometric Template Update via Feature Alignment Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Fingerprint sensors face degraded biometric performance due to fixed patterns, which can lead to a higher false accept rate and vulnerability after the enrollment phase, as these patterns can be introduced into the biometric template, compromising security.
Innovation Solution
A method that considers the alignment of features when comparing biometric measurements, identifying and removing features with excessive alignment as potential fixed patterns, allowing for dynamic template updates and improved security by discarding only features with alignment differences below a threshold.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If fixed pattern mitigation is implemented by comparing alignment of features, then false accept rate is reduced and security is improved, but device complexity increases due to additional alignment analysis requirements
Solution Approach 1:
The feature set is segmented into genuine features and fixed pattern features based on alignment characteristics. By dividing the feature comparison into alignment analysis and matching stages, the system identifies and excludes features that are too well-aligned (indicative of fixed patterns) while processing only relevant features for authentication, thus improving security without requiring complete rejection of measurements.
Solution Approach 2:
The system applies different processing quality to different features based on their alignment properties. Features with excessive alignment are treated differently (identified as fixed patterns and excluded) compared to features with normal alignment variations. This localized differentiation allows precise mitigation of fixed patterns while preserving genuine biometric information.
2Reliability
If subsequent biometric measurements are discarded due to small differences, then fixed pattern detection is simplified, but loss of information increases and usability decreases
Solution Approach 1:
Instead of completely discarding measurements with small differences (excessive action), the system applies partial action by selectively identifying and excluding only the specific features that exhibit fixed pattern characteristics (excessive alignment). This partial filtering approach retains useful biometric information while removing harmful fixed pattern components, thus reducing information loss and maintaining usability.
Solution Approach 2:
The measurement data is segmented into useful biometric features and harmful fixed pattern features. By dividing the feature set and applying different processing rules to each segment, the system preserves genuine biometric information while excluding fixed patterns, thereby minimizing information loss while achieving reliable fixed pattern detection.
3Measurement precision
If feature alignment comparison is performed to identify fixed patterns, then measurement precision is improved, but difficulty of detecting and measuring increases due to additional alignment analysis
Solution Approach 1:
Alignment comparison is performed as a preliminary action before final feature matching. By pre-identifying features with excessive alignment (fixed patterns) and excluding them beforehand, the system simplifies the subsequent matching process. This preliminary filtering reduces the complexity of the main detection task while maintaining high measurement precision through the alignment analysis.
Data Source
Figure 1~2
Figure 3a~3f
Figure 4~5
AI summary
The present invention generally relates to a method for analyzing biometric measurements of a user, and specifically to determination of an alignment level between features of different biometric measurements acquired using a biometric sensor. The invention also relates to a corresponding fingerprint sensing system and to a computer program product.