Encrypted Biometric Matching via Group Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing biometric authentication systems face challenges in executing one-to-N authentication efficiently due to high calculation costs and the difficulty of performing matching processing that permits ambiguity in information, especially when biological information is concealed for security reasons.
Innovation Solution
An encrypted information matching system that classifies information into groups based on similarity and uses representative information to determine the group for input data, allowing for secure and efficient matching by calculating similarities between encrypted input and stored information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If one-to-N biometric authentication is executed by comparing input data with all stored templates, then authentication accuracy is improved, but calculation cost and processing time increase significantly
Solution Approach 1:
The patent divides the large-scale template database into multiple groups based on feature values. The identification unit performs segmentation by comparing the input feature value with group information to identify a candidate group, then only performs detailed matching within that group. This segmentation reduces the number of comparisons from N (total templates) to M (templates in candidate group), where M << N, thereby improving processing speed while maintaining authentication accuracy.
Solution Approach 2:
The patent performs preliminary grouping of templates based on feature values before the actual authentication process. The grouping information is stored in advance, allowing the identification unit to quickly narrow down the search space. This preliminary action of organizing data into groups enables the system to skip unnecessary comparisons and directly focus on relevant templates, significantly reducing calculation cost while ensuring accurate authentication.
2Reliability
If biometric information is encrypted for security, then confidentiality is improved, but the ability to perform matching processing that permits ambiguity deteriorates
Solution Approach 1:
The patent changes the parameter representation by using encrypted feature values and encrypted group information instead of plaintext data. The identification unit performs calculations on encrypted parameters, comparing the input encrypted feature value with encrypted group information to identify candidate groups. This parameter transformation allows matching processing to be performed on encrypted data, maintaining confidentiality while enabling ambiguity-permitting operations through encrypted comparison.
Solution Approach 2:
The patent introduces encrypted group information as an intermediary that facilitates matching between encrypted input data and encrypted template data. The identification unit uses this intermediary to perform preliminary filtering without decrypting the actual biometric data. The encrypted group information acts as a mediator that enables the system to narrow down candidate templates while maintaining the encrypted state, thus preserving both confidentiality and matching capability.
3Reliability
If all stored templates are used as matching targets, then authentication security is improved, but calculation load and processing time increase
Solution Approach 1:
The patent segments the template database into multiple groups based on feature values and stores group information in advance. During authentication, the identification unit performs segmentation by comparing input data with group information to identify a candidate group, then restricts detailed matching to only templates within that group. This segmentation reduces the calculation load from comparing against all N templates to comparing against M templates in the candidate group, where M << N, thereby reducing energy consumption while maintaining security through comprehensive matching within the relevant subset.
Solution Approach 2:
The patent performs preliminary organization of templates into groups based on feature values before authentication. This preliminary action creates an efficient search structure that allows the system to quickly identify candidate groups without performing full comparisons. By pre-processing the data into grouped structures, the system reduces real-time calculation load during authentication while ensuring that security is maintained through subsequent detailed matching within the identified candidate groups.
Data Source
AI summary
An encrypted information matching device 20 includes: a storage unit 240 in which representative information representing a group based on similarity in a characteristic and encrypted information to be matched belonging to the group are associated with each other and stored; an identification unit 231 which calculates a first similarity between the encrypted input information and the representative information stored in the unit 240, and which, based on the first similarity, identifies a group to be matched with the first input information; and a determination unit 232 which calculates a second similarity between the encrypted input information and the encrypted information to be matched that belongs to the group identified, in a state in which the input information and the information to be matched are encrypted, wherein the unit 232 determines the presence or absence of information to be matched which makes the calculated second similarity satisfy criteria.