Privacy-Enabled Biometric Processing via Homomorphic Encryption
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional biometric systems face limitations in performing one to many searches and matching on encrypted biometric information, especially in polynomial time, while also dealing with key management overhead and security vulnerabilities.
Innovation Solution
The implementation of a privacy-enabled biometric system that uses deep neural networks for one-way homomorphic encryption, allowing for computations and comparisons on encrypted biometric data without decryption, thereby enhancing security and scalability.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional biometric systems store and search unencrypted biometric data, then search and matching operations can be performed efficiently, but security and privacy of biometric data are compromised
Solution Approach 1:
The patent introduces homomorphic encryption as an intermediary mechanism that allows search operations to be performed on encrypted biometric data without decryption. The encrypted feature vectors serve as a mediator between the need for security (keeping data encrypted) and the need for efficient searching (performing operations without full decryption), enabling polynomial-time one-to-many matching while maintaining data confidentiality
Solution Approach 2:
The system transforms biometric data into a different parameter space through deep neural network processing, converting raw biometric inputs into encrypted feature vectors. This parameter transformation enables efficient Euclidean distance-based searching in the encrypted space, resolving the contradiction between maintaining security through encryption and achieving efficient search operations
2Productivity
If one-to-many searching is performed on encrypted biometric data using conventional methods, then security is maintained, but the search cannot be completed in polynomial time
Solution Approach 1:
The patent replaces conventional decryption-based search mechanics with homomorphic encryption-based computation. Instead of decrypting data to perform searches (mechanical decryption process), the system uses mathematical properties of homomorphic encryption to perform distance calculations directly on encrypted feature vectors, achieving polynomial-time complexity while maintaining encryption
Solution Approach 2:
The system performs preliminary transformation of biometric data into encrypted feature vectors using deep neural networks before storage. This preliminary action enables subsequent efficient searching operations to be performed directly on the pre-processed encrypted representations without requiring additional decryption steps
3Reliability
If biometric data is encrypted using conventional encryption methods, then security is improved, but key management overhead increases significantly
Solution Approach 1:
The patent extracts the key management burden from the biometric processing system by using homomorphic encryption properties that allow computations on encrypted data without requiring the decryption key. The secret key is taken out of the operational flow, eliminating the need for complex key management infrastructure while maintaining strong security
Solution Approach 2:
The homomorphic encryption scheme enables the encrypted biometric data to serve itself for search operations without external key management intervention. The encrypted feature vectors can be searched and compared directly using their mathematical properties, making the system self-sufficient and eliminating dependency on complex key management systems
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.