Encrypted Biometric Matching Using Homomorphic Feature Vectors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional biometric systems face limitations in performing one-to-many searches on encrypted biometric data, leading to security vulnerabilities and inefficiencies, particularly in key management and scalability.
Innovation Solution
A privacy-enabled biometric system utilizing deep neural networks (DNNs) processes encrypted biometric feature vectors through one-way homomorphic encryption, enabling secure and efficient one-to-many matching and classification without decrypting the original data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional biometric systems store and search biometric data in clear text, then search efficiency and matching speed are improved, but security and privacy of biometric data are compromised
Solution Approach 1:
The patent introduces homomorphic encryption as an intermediary mechanism that enables search operations on encrypted biometric data without decryption. The encrypted feature vectors serve as mediators between the stored biometric data and search queries, allowing the system to perform matching operations while maintaining data encryption throughout the process.
Solution Approach 2:
The system transforms biometric data into encrypted feature vectors using homomorphic encryption, changing the parameter state from plaintext to ciphertext. This parameter transformation allows the data to maintain its usability for search operations while achieving the desired security properties through encryption.
2Reliability
If biometric data is encrypted using conventional methods, then security is improved, but the ability to perform one-to-many searches and matching operations is lost
Solution Approach 1:
Homomorphic encryption serves as an intermediary that preserves the structural properties of biometric data in encrypted form, enabling search operations. The encrypted feature vectors maintain enough mathematical structure to allow comparison and matching operations without requiring decryption, thus preserving search capability while achieving security.
Solution Approach 2:
The patent replaces conventional decryption-based search mechanisms with homomorphic encryption-based operations. Instead of decrypting data to perform searches and then re-encrypting results, the system uses homomorphic properties to perform matching operations directly on encrypted data, substituting the mechanical decryption-encryption cycle with a more efficient cryptographic operation.
3Reliability
If key management is implemented to secure stored biometric data, then security is improved, but system complexity and overhead increase
Solution Approach 1:
The patent extracts the decryption operation from the search process entirely. By using homomorphic encryption, the system removes the need to manage decryption keys for search operations, as all matching can be performed on encrypted data. This extraction of the decryption step eliminates a major source of key management complexity.
Solution Approach 2:
The system works with encrypted copies of biometric data throughout the entire process. Instead of requiring access to original plaintext data or managing keys to access it, the system creates and operates on encrypted representations, eliminating the need for plaintext key management while maintaining full functional capability.
4Device complexity
If biometric matching is limited to one-to-one comparisons, then security is simplified, but scalability and productivity are reduced
Solution Approach 1:
The homomorphic encryption approach provides a universal mechanism that handles both one-to-one and one-to-many matching operations through the same cryptographic framework. The encrypted feature vector comparison operation serves multiple functions, enabling scalable database searches without requiring separate mechanisms for different matching types, thus achieving both simplicity and scalability.
Data Source
AI summary
In one embodiment, a set of feature vectors can be derived from any biometric data, and then using a deep neural network (“DNN”) on those one-way homomorphic encryptions (i.e., each biometrics' feature vector) can determine matches or execute searches on encrypted data. Each biometrics' feature vector can then be stored and/or used in conjunction with respective classifications, for use in subsequent comparisons without fear of compromising the original biometric data. In various embodiments, the original biometric data is discarded responsive to generating the encrypted values. In another embodiment, the homomorphic encryption enables computations and comparisons on cypher text without decryption. This improves security over conventional approaches. Searching biometrics in the clear on any system, represents a significant security vulnerability. In various examples described herein, only the one-way encrypted biometric data is available on a given device. Various embodiments restrict execution to occur on encrypted biometrics for any matching or searching.