Homomorphic Encryption for Speaker Recognition: Protection of Biometric Templates and Vendor Model Parameters

@article{Nautsch2018HomomorphicEF,
  title={Homomorphic Encryption for Speaker Recognition: Protection of Biometric Templates and Vendor Model Parameters},
  author={Andreas Nautsch and Sergey Isadskiy and Jascha Kolberg and Marta Gomez-Barrero and Christoph Busch},
  journal={ArXiv},
  year={2018},
  volume={abs/1803.03559}
}
Data privacy is crucial when dealing with biometric data. Accounting for the latest European data privacy regulation and payment service directive, biometric template protection is essential for any commercial application. Ensuring unlinkability across biometric service operators, irreversibility of leaked encrypted templates, and renewability of e.g., voice models following the i-vector paradigm, biometric voice-based systems are prepared for the latest EU data privacy legislation. Employing… 

Figures and Tables from this paper

Preserving privacy in speaker and speech characterisation
Secure Fingerprint Authentication with Homomorphic Encryption
TLDR
This paper proposes a homomorphic encryption-based fingerprint authentication system to provide access control, while protecting sensitive biometric template data, and studies the trade-off between the computational overload and authentication accuracy.
Privacy-Preserving Speaker Recognition with Cohort Score Normalisation
TLDR
This paper proposes the first computationally feasible approach to privacy-preserving cohort score normalisation using probabilistic linear discriminant analysis (PLDA) comparisons, based on a cohort pruning scheme based on secure multi-party computation.
Privacy-preserving speaker verification system based on binary I-vectors
European Union's Horizon 2020 research and innovation program, Grant/Award Numbers: 653586, 769872 Abstract Speaker verification is a key technology in many services and applications like smartphones
Privacy-preserving speech processing via STPC and TEEs (Poster)
TLDR
VoiceGuard is presented, an architecture that efficiently protects speech processing inside a trusted execution environment (TEE) and Offline Model Guard (OMG) is presented to enable privacypreserving speech processing on the predominant mobile computing platform ARM even in offline scenarios.
Towards End-to-End Private Automatic Speaker Recognition
TLDR
This work shows how speaker embeddings can be extracted while keeping both the speaker’s voice and the service provider's model private, using Secure Multiparty Computation, and shows that it is possible to obtain reasonable trade-offs between security and computational cost.
Privacy-Preserving Voice Anti-Spoofing Using Secure Multi-Party Computation
TLDR
The first solution based upon the combination of shallow neural networks with secure MPC is reported, which is not only computationally efficient, but that it also improves upon the performance of the ASVspoof baseline countermeasure, all while preserving privacy.
Security, Privacy, and Usability in Continuous Authentication: A Survey
TLDR
An overview of continuous authentication is presented and issues that these modes have related to security, privacy, and usability are discussed, and a comparison of privacy-preserving approaches dealing with the privacy issues is provided.
Audio-Visual Biometric Recognition and Presentation Attack Detection: A Comprehensive Survey
TLDR
A comprehensive survey on existing state-of-the-art audio-visual recognition techniques, publicly available databases for benchmarking, and Presentation Attack Detection (PAD) algorithms is made.
...
...

References

SHOWING 1-10 OF 48 REFERENCES
Multi-biometric template protection based on Homomorphic Encryption
Privacy Protection in Biometric-Based Recognition Systems: A marriage between cryptography and signal processing
Systems employing biometric traits for people authentication and identification are witnessing growing popularity due to the unique and indissoluble link between any individual and his/her biometric
Implementation of Fixed-Length Template Protection Based on Homomorphic Encryption with Application to Signature Biometrics
TLDR
This work proposes a new effcient biometric template protection scheme based on homomorphic probabilistic encryption for fixed-length templates, where only encrypted data is handled and is robust to hill-climbing and inverse-biometrics attacks.
Privacy-Preserving Comparison of Variable-Length Data With Application to Biometric Template Protection
TLDR
A new approach to efficiently compare variable-length data in the encrypted domain using homomorphic encryption where only encrypted data is stored or exchanged and is fused with existing fixed-length techniques in order to obtain increased comparison accuracy.
Multi-Bit Allocation: Preparing Voice Biometrics for Template Protection
TLDR
This paper introduces a binarization technique for voice biometric features called multi-bit allocation and shows that the resulting binary representation causes only a marginal decrease in biometric performance compared to the baseline system, confirming the soundness and aplicability of the proposed scheme.
Secure and Efficient Protocols for Iris and Fingerprint Identification
TLDR
This work develops and implements the first privacy-preserving identification protocol for iris codes and designs and implements a secure protocol for fingerprint identification based on FingerCodes with a substantial improvement in the performance compared to existing solutions.
A privacy-compliant fingerprint recognition system based on homomorphic encryption and Fingercode templates
  • M. Barni, T. Bianchi, F. Scotti
  • Computer Science
    2010 Fourth IEEE International Conference on Biometrics: Theory, Applications and Systems (BTAS)
  • 2010
TLDR
A novel complete demonstrator based on a distributed biométrie system that is capable to protect the privacy of the individuals by exploiting cryptosystems and computes the matching task in the encrypted domain by exploiting homomorphic encryption and using Fingercode templates.
Anonymous Biometric Access Control
TLDR
The -Anonymous Quantization (kAQ) framework is proposed that provides an effective and secure tradeoff of privacy and complexity and demonstrates the validity of the framework and illustrate a practical implementation of an anonymous biometric system.
Privacy-Preserving Biometric Authentication and Matching via Lattice-Based Encryption
TLDR
This work exploits the homomorphic properties of NTRU to provide a robust and secure solution for privacy-preserving biometric authentication using lattice-based encryption and provides experimental results which illustrate the efficacy of this proposal.
...
...