Biometric Search Filtering Using Dynamic Subset Generation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current biometric search methods in multi-biometric databases are inefficient due to the high time and computational requirements of 1:N searches, as existing filters are modality-specific and require a priori classification, limiting their applicability across different biometric modalities.
Innovation Solution
A system and method that generates a subset of enrolled biometric representations using pre-computed vectors derived from a filter set, allowing for dynamic filtering and reducing search time by comparing a probe biometric against a dynamically generated subset of candidates, applicable across various biometric modalities.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a 1:N search is performed against a master database of enrolled biometric representations, then search completeness is maintained, but search time and computational resources increase prohibitively
Solution Approach 1:
The patent segments the master database into multiple subsets based on extracted features (such as gender, age group, or other distinguishable characteristics). Instead of searching the entire database, the system divides the search space into manageable segments and only searches relevant subsets, significantly reducing search time while maintaining completeness through systematic coverage of all segments.
Solution Approach 2:
The patent performs preliminary extraction and comparison of modality-specific features before the main biometric matching process. By pre-computing and storing feature representations (such as ridge flow patterns, core-delta locations, or other modality-specific characteristics), the system enables rapid filtering during search time without requiring full biometric comparisons for all entries, thus reducing computational load and search time.
2Productivity
If modality-specific filters are used to reduce search time, then search speed improves, but adaptability across multi-biometric databases deteriorates
Solution Approach 1:
The patent creates a universal filtering mechanism that can operate across multiple biometric modalities (fingerprints, iris, facial recognition, etc.) by extracting and comparing modality-specific features in a standardized manner. The system maintains a master database that can accommodate different biometric types while using a common feature extraction and comparison framework, enabling the same search algorithm to efficiently handle various modalities without requiring modality-specific search procedures.
3Device complexity
If pre-assignment of biometrics to bins is performed, then search organization improves, but search accuracy deteriorates due to dependence on initial binning decision accuracy
Solution Approach 1:
The patent employs dynamic filtering that adapts to the specific characteristics of each probe biometric during the search process. Rather than relying on static pre-defined bins, the system dynamically extracts modality-specific features from the probe and uses these features to dynamically determine which database entries deserve further consideration. This dynamic approach allows the search to adapt to variations in biometric data quality, presentation conditions, and individual characteristics, improving search accuracy while maintaining organized database structure.
Data Source
AI summary
A system and method for enabling analysis of enrolled biometric data is presented. A plurality of vectors each having a plurality of score values representative of the relationship between individual ones of the enrolled biometrics with a plurality of biometric representations forming a filter set are described. Judicious use of the vectors enables a filtering of the enrolled biometric data on a dynamic basis.