Design Choices for X-Vector Based Speaker Anonymization

Srivastava, Brij Mohan Lal; Tomashenko, Natalia; Wang, Xin; Yamagishi, Junichi; Maouche, Mohamed; Bellet, Aurélien; Tommasi, Marc

doi:10.21437/interspeech.2020-2692

Cited by 26 publications

(17 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The experiments done in [2,14] showed that black and grey-box attackers can easily be fooled when enrollment speech and trial speech are anonymized in a different manner, i.e. by selecting the x-vector in different regions, applying the F0 linear transformation and random gender selection.…”

Section: Discussionmentioning

confidence: 99%

“…To select the target pseudo-speaker identity, module B (from Figure 1) has many hyper-parameters that affect the selection mechanism. According to [2,13,14], the best anonymization results are achieved by picking the pseudo-speaker in a dense region of the x-vector space, randomly targeting male or female gender pseudo-speaker, and modifying the F0 values of the input speech so that it matches the F0 statistics of the real speakers used to generate the pseudoidentity.…”

Section: Design Choices For Anonymizationmentioning

confidence: 99%

See 1 more Smart Citation

Evaluating X-Vector-Based Speaker Anonymization Under White-Box Assessment

Champion

Jouvet

Larcher

2021

Speech and Computer

View full text Add to dashboard Cite

In the scenario of the Voice Privacy challenge, anonymization is achieved by converting all utterances from a source speaker to match the same target identity; this identity being randomly selected. In this context, an attacker with maximum knowledge about the anonymization system can not infer the target identity. This article proposed to constrain the target selection to a specific identity, i.e., removing the random selection of identity, to evaluate the extreme threat under a whitebox assessment (the attacker has complete knowledge about the system). Targeting a unique identity also allows us to investigate whether some target's identities are better than others to anonymize a given speaker.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Design Choices For Anonymizationmentioning

confidence: 99%

Evaluating X-Vector-Based Speaker Anonymization Under White-Box Assessment

Champion

Jouvet

Larcher

2021

Speech and Computer

View full text Add to dashboard Cite

show abstract

“…speaker identity) while leaving linguistic content intact. Most of the proposed works focus on protecting/anonymizing speaker identity using voice conversion (VC) mechanisms [8,9,10]. These VC methods, however, aim to protect the speaker identity against different linkage attacks limited by the attacker's knowledge [11].…”

Section: Privacy-preserving Voice Analyticsmentioning

confidence: 99%

Configurable Privacy-Preserving Automatic Speech Recognition

Aloufi¹,

Haddadi²,

Boyle³

2021

Interspeech 2021

View full text Add to dashboard Cite

Voice assistive technologies have given rise to far-reaching privacy and security concerns. In this paper we investigate whether modular automatic speech recognition (ASR) can improve privacy in voice assistive systems by combining independently trained separation, recognition, and discretization modules to design configurable privacy-preserving ASR systems. We evaluate privacy concerns and the effects of applying various stateof-the-art techniques at each stage of the system, and report results using task-specific metrics (i.e. WER, ABX, and accuracy). We show that overlapping speech inputs to ASR systems present further privacy concerns, and how these may be mitigated using speech separation and optimization techniques. Our discretization module is shown to minimize paralinguistics privacy leakage from ASR acoustic models to levels commensurate with random guessing. We show that voice privacy can be configurable, and argue this presents new opportunities for privacy-preserving applications incorporating ASR.

show abstract

“…Finally, the original linguistic features, pitch, and anonymised x-vectors were used to synthesise anonymised speech by means of neural acoustic and waveform models. More details on B1 development are available in [2,3,11].…”

Section: Anonymisation Systemsmentioning

confidence: 99%

Anonymous Speaker Clusters: Making Distinctions Between Anonymised Speech Recordings with Clustering Interface

O’Brien¹,

Tomashenko²,

Chanclu³

et al. 2021

Interspeech 2021

View full text Add to dashboard Cite

Our study examined the performance of evaluators tasked to group natural and anonymised speech recordings into clusters based on their perceived similarities. Speech stimuli were selected from the VCTK corpus; two systems developed for the VoicePrivacy 2020 Challenge were used for anonymisation. The Baseline-1 (B1) system was developed by using x-vectors and neural waveform models, while the Baseline-2 (B2) system relied on digital-signal-processing techniques. 74 evaluators completed three trials composed of 16 recordings with either natural or anonymised speech generated from a single system. F-measure and cluster purity metrics were used to assess evaluator accuracy. Probabilistic linear discriminant analysis (PLDA) scores from an automatic speaker verification system were generated to quantify similarity between recordings and used to correlate subjective results. Our findings showed that non-native English speaking evaluators significantly lowered their F-measure means when presented anonymised recordings. We observed no significance for cluster purity. Pearson correlation procedures revealed that PLDA scores generated from natural and B2-anonymised speech recordings correlated positively to F-measure and cluster purity metrics. These findings show evaluators were able to use the interface to cluster natural and anonymised speech recordings and suggest anonymisation systems modelled like B1 are more effective at suppressing identifiable speech characteristics.

show abstract

Design Choices for X-Vector Based Speaker Anonymization

Cited by 26 publications

References 0 publications

Evaluating X-Vector-Based Speaker Anonymization Under White-Box Assessment

Evaluating X-Vector-Based Speaker Anonymization Under White-Box Assessment

Configurable Privacy-Preserving Automatic Speech Recognition

Anonymous Speaker Clusters: Making Distinctions Between Anonymised Speech Recordings with Clustering Interface

Contact Info

Product

Resources

About