Fuzzy Vault Encryption for Multi-Item Biometric Database Search
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing 'fuzzy vault' scheme for data encryption is limited to comparing two data sets and cannot be applied to comparing data with a set of multiple data items in a database, which is necessary for biometric identification applications.
Innovation Solution
A method of encrypting a set of indexed data by generating random encoding functions for each datum, creating a protected set by adding code word coordinates and error-inducing points, and reindexing elements to secure the data, allowing for comparison with a set of data without revealing information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the fuzzy vault scheme is used to encrypt data, then the security and privacy of the data are improved, but the scheme is limited to comparing only two data sets and cannot be applied to comparing data with a set of multiple data items in a database
Solution Approach 1:
The invention segments the database into multiple protected sets, where each protected set contains encrypted data for a specific data item. Each protected set is constructed independently using the fuzzy vault scheme, allowing the system to maintain the security properties of the original scheme while enabling comparison across multiple data items. The segmentation is achieved by creating separate protected sets P1, P2, ..., Pn for each data item D1, D2, ..., Dn in the database.
Solution Approach 2:
The invention extends the fuzzy vault scheme to provide universal functionality for comparing a test data set against multiple database entries simultaneously. The decryption algorithm is enhanced to accept a test data set B and compare it against all protected sets in the database, returning all matching data items that exceed a similarity threshold. This multi-functional capability allows the same encryption framework to serve both pairwise comparison and multi-item database search functions.
2Reliability
If error inducing points are added to mask the information related to the data, then the security against unauthorized access is improved, but the complexity of the encryption process and the size of the protected set increase
Solution Approach 1:
The invention adjusts the parameters of the fuzzy vault scheme, specifically the number of error-inducing points E and the similarity threshold T, to optimize the balance between security and complexity. By carefully selecting these parameters, the system achieves adequate security protection while controlling the size of the protected sets and the computational complexity of the encryption and decryption processes. The parameters can be tuned based on the specific security requirements and performance constraints of the application.
Data Source
Figure 1a
Figure 1b
Figure 2
AI summary
The invention concerns a server-implemented method for encrypting at least two pieces of indexed data, the data being in the form of lists of elements, each element of which belongs to a finite set of indexed symbols called an alphabet, the method being characterised in that the data is encrypted to form a protected set, the step of encrypting and creating the protected set comprising the following steps: - the server randomly generates, for each datum in the base, a corresponding encoding function, and -for each symbol of the alphabet, .for each datum, .if at least one element of the list that constitutes the datum is the symbol of the alphabet, the server determines the image of said symbol of the alphabet via the encoding function corresponding to the datum in order to obtain a codeword coordinate that is a function of the datum and of the symbol of the alphabet, .the server adds the codeword coordinate thereby obtained to an indexed set corresponding to the element of the alphabet, the set having a predetermined cardinal, . then, the server completes the indexed set corresponding to the element of the alphabet with error-inducing points, . the server randomly reindexes the elements of the indexed set corresponding to the symbol of the alphabet, and . the server adds the indexed set corresponding to the symbol of the alphabet to the protected set. The invention also concerns a method for identifying an individual, implemented on data encrypted by the encryption method according to the invention.