Privacy and Utility of X-Vector Based Speaker Anonymization

We study the scenario where individuals ( speakers ) contribute to the publication of an anonymized speech corpus. Data users leverage this public corpus for downstream tasks, e.g., training an automatic speech recognition (ASR) system, while attackers may attempt to de-anonymize it using auxiliary...

Ausführliche Beschreibung

Gespeichert in:

Bibliographische Detailangaben
Veröffentlicht in:	IEEE/ACM transactions on audio, speech, and language processing Jg. 30; S. 2383 - 2395
Hauptverfasser:	Srivastava, Brij Mohan Lal, Maouche, Mohamed, Sahidullah, Md, Vincent, Emmanuel, Bellet, Aurelien, Tommasi, Marc, Tomashenko, Natalia, Wang, Xin, Yamagishi, Junichi
Format:	Journal Article
Sprache:	Englisch
Veröffentlicht:	Piscataway IEEE 01.01.2022 The Institute of Electrical and Electronics Engineers, Inc. (IEEE) Institute of Electrical and Electronics Engineers
Schlagworte:	Artificial Intelligence Audio equipment Automatic speech recognition Cloning Computation and Language Computer Science Corpus linguistics Error analysis Feature extraction Linkability Machine Learning Measurement Privacy Quality assessment speaker anonymization Speech recognition Speech synthesis Task analysis Training voice conversion Voice recognition linkability privacy utility speech recognition speaker identification speaker anonymization
ISSN:	2329-9290, 2329-9304
Online-Zugang:	Volltext
Tags:	Tag hinzufügen Keine Tags, Fügen Sie den ersten Tag hinzu!

Abstract	We study the scenario where individuals ( speakers ) contribute to the publication of an anonymized speech corpus. Data users leverage this public corpus for downstream tasks, e.g., training an automatic speech recognition (ASR) system, while attackers may attempt to de-anonymize it using auxiliary knowledge. Motivated by this scenario, speaker anonymization aims to conceal speaker identity while preserving the quality and usefulness of speech data. In this article, we study x-vector based speaker anonymization, the leading approach in the VoicePrivacy Challenge, which converts the speaker's voice into that of a random pseudo-speaker. We show that the strength of anonymization varies significantly depending on how the pseudo-speaker is chosen. We explore four design choices for this step: the distance metric between speakers, the region of speaker space where the pseudo-speaker is picked, its gender, and whether to assign it to one or all utterances of the original speaker. We assess the quality of anonymization from the perspective of the three actors involved in our threat model, namely the speaker, the user and the attacker. To measure privacy and utility, we use respectively the linkability score achieved by the attackers and the decoding word error rate achieved by an ASR model trained on the anonymized data. Experiments on LibriSpeech show that the best combination of design choices yields state-of-the-art performance in terms of both privacy and utility. Experiments on Mozilla Common Voice further show that it guarantees the same anonymization level against re-identification attacks among 50 speakers as original speech among 20,000 speakers.
AbstractList	We study the scenario where individuals ( speakers ) contribute to the publication of an anonymized speech corpus. Data users leverage this public corpus for downstream tasks, e.g., training an automatic speech recognition (ASR) system, while attackers may attempt to de-anonymize it using auxiliary knowledge. Motivated by this scenario, speaker anonymization aims to conceal speaker identity while preserving the quality and usefulness of speech data. In this article, we study x-vector based speaker anonymization, the leading approach in the VoicePrivacy Challenge, which converts the speaker's voice into that of a random pseudo-speaker. We show that the strength of anonymization varies significantly depending on how the pseudo-speaker is chosen. We explore four design choices for this step: the distance metric between speakers, the region of speaker space where the pseudo-speaker is picked, its gender, and whether to assign it to one or all utterances of the original speaker. We assess the quality of anonymization from the perspective of the three actors involved in our threat model, namely the speaker, the user and the attacker. To measure privacy and utility, we use respectively the linkability score achieved by the attackers and the decoding word error rate achieved by an ASR model trained on the anonymized data. Experiments on LibriSpeech show that the best combination of design choices yields state-of-the-art performance in terms of both privacy and utility. Experiments on Mozilla Common Voice further show that it guarantees the same anonymization level against re-identification attacks among 50 speakers as original speech among 20,000 speakers. We study the scenario where individuals (speakers) contribute to the publication of an anonymized speech corpus. Data users then leverage this public corpus to perform downstream tasks (such as training automatic speech recognition systems), while attackers may try to de-anonymize itbased on auxiliary knowledge they collect. Motivated by this scenario, speaker anonymization aims to conceal the speaker identity while preserving the quality and usefulness of speech data. In this paper, we study x-vector based speaker anonymization, the leading approach in the recent Voice Privacy Challenge, which converts an input utterance into that of a random pseudo-speaker. We show that the strength of the anonymization varies significantly depending on how the pseudo-speaker is selected. In particular, we investigate four design choices: the distance measure between speakers, the region of x-vector space where the pseudo-speaker is mapped, the gender selection and whether to use speaker or utterance level assignment. We assess the quality of anonymization from the perspective of the three actors involved in our threat model, namely the speaker, the user and the attacker. To measure privacy and utility, we use respectively the linkability score achieved by the attackers and the decoding word error rate incurred by an ASR model trained with the anonymized data. Experiments on LibriSpeech dataset confirm that the optimal combination ofdesign choices yield state-of-the-art performance in terms of privacy protection as well as utility. Experiments on Mozilla Common Voice dataset show that the best design choices with 50 speakers guarantee the same anonymization level against re-identification attack as raw speech with 20,000 speakers.
Author	Sahidullah, Md Srivastava, Brij Mohan Lal Wang, Xin Vincent, Emmanuel Maouche, Mohamed Tomashenko, Natalia Yamagishi, Junichi Bellet, Aurelien Tommasi, Marc
Author_xml	– sequence: 1 givenname: Brij Mohan Lal orcidid: 0000-0001-7531-7123 surname: Srivastava fullname: Srivastava, Brij Mohan Lal email: brij.srivastava@inria.fr organization: Inria, CNRS, Centrale, Univ. Lille, Lille, France – sequence: 2 givenname: Mohamed surname: Maouche fullname: Maouche, Mohamed email: mohamed.maouche@inria.fr organization: Inria, CNRS, Centrale, Univ. Lille, Lille, France – sequence: 3 givenname: Md orcidid: 0000-0002-0624-2903 surname: Sahidullah fullname: Sahidullah, Md email: md.sahidullah@inria.fr organization: Université de Lorraine, CNRS, Inria, LORIA, Nancy, France – sequence: 4 givenname: Emmanuel orcidid: 0000-0002-0183-7289 surname: Vincent fullname: Vincent, Emmanuel email: emmanuel.vincent@inria.fr organization: Université de Lorraine, CNRS, Inria, LORIA, Nancy, France – sequence: 5 givenname: Aurelien surname: Bellet fullname: Bellet, Aurelien email: aurelien.bellet@inria.fr organization: Inria, CNRS, Centrale, Univ. Lille, Lille, France – sequence: 6 givenname: Marc surname: Tommasi fullname: Tommasi, Marc email: marc.tommasi@inria.fr organization: Inria, CNRS, Centrale, Univ. Lille, Lille, France – sequence: 7 givenname: Natalia surname: Tomashenko fullname: Tomashenko, Natalia email: natalia.tomashenko@univ-avignon.fr organization: Laboratoire Informatique d' Avignon (LIA), Avignon Université, Avignon, France – sequence: 8 givenname: Xin orcidid: 0000-0001-8246-0606 surname: Wang fullname: Wang, Xin email: wangxin@nii.ac.jp organization: National Institute of Informatics, Tokyo, Japan – sequence: 9 givenname: Junichi orcidid: 0000-0003-2752-3955 surname: Yamagishi fullname: Yamagishi, Junichi email: jyamagis@nii.ac.jp organization: National Institute of Informatics, Tokyo, Japan
BackLink	https://inria.hal.science/hal-03197376$$DView record in HAL
BookMark	eNp9kD9PwzAQxS1UJErpF4AlEhNDytnOP4upVECRIlGpLWKzbMcRLmlcnLRS-PSkDXRgYLrT6b27d79z1CttqRG6xDDCGNjtYjxPZyMChIwoZhAH-AT1CSXMZxSC3m9PGJyhYVWtAABDzFgc9NHdzJmdUI0nysxb1qYwdePZ3HvzX7WqrfPuRaUzb77R4kM7b9yebtbmS9TGlhfoNBdFpYc_dYCWjw-LydRPX56eJ-PUVwFltR9mmOQaQh3LKAqVSrCKQynDMEtkAIpJCTiXOhcilxmlmGhKQixZIiEII8joAN10e99FwTfOrIVruBWGT8cp38-gfTumcbSjrfa6026c_dzqquYru3VlG4-TiEUkCaIgalWkUylnq8rp_LgWA98z5QemfM-U_zBtTckfkzL1AUTthCn-t151VqO1Pt5iCWFtIPoN6tOErQ
CODEN	ITASFA
CitedBy_id	crossref_primary_10_1109_TASLP_2024_3492793 crossref_primary_10_1109_TIFS_2025_3556345 crossref_primary_10_1016_j_csl_2025_101810 crossref_primary_10_1109_TASLP_2024_3407600 crossref_primary_10_1109_ACCESS_2025_3546618 crossref_primary_10_1109_TASLP_2023_3313429 crossref_primary_10_3390_app14093876 crossref_primary_10_1109_TIFS_2025_3592553 crossref_primary_10_1038_s43856_024_00609_5 crossref_primary_10_1109_JPROC_2023_3276209 crossref_primary_10_1016_j_patrec_2024_10_016 crossref_primary_10_1177_1088467X241301384 crossref_primary_10_1109_TASLP_2024_3430530
Cites_doi	10.1109/ASRU.2009.5373356 10.21437/Interspeech.2020-1887 10.21437/Interspeech.2018-1929 10.1109/ICASSP.2016.7472923 10.1109/ICME46284.2020.9102875 10.1016/j.csl.2022.101362 10.1109/FG.2015.7285021 10.1007/11744085_41 10.1109/ICASSP40776.2020.9053868 10.1561/9781601988195 10.1016/0167-6393(95)00009-D 10.1109/ICASSP.2014.6853882 10.21437/Interspeech.2020-2692 10.21437/Odyssey.2016-59 10.1109/ASRU46091.2019.9003979 10.1016/j.csl.2019.05.005 10.1007/978-3-540-71050-9 10.1016/j.csl.2017.05.001 10.1109/ICASSP.2016.7472729 10.1109/ICASSP.2018.8461375 10.21437/Interspeech.2008-644 10.1016/j.csl.2019.101024 10.21437/Interspeech.2019-2441 10.21437/Interspeech.2019-2415 10.21437/Interspeech.2020-1333 10.1109/FSKD.2007.347 10.21437/Odyssey.2018-36 10.21437/Interspeech.2011-91 10.1109/ICASSP39728.2021.9413386 10.1016/j.csl.2005.08.001 10.1109/ICASSP.2019.8682298 10.21437/Interspeech.2021-1070 10.21437/Interspeech.2019-1572 10.21437/Interspeech.2017-950 10.21437/Interspeech.2020-1090 10.1016/j.csl.2022.101351 10.1109/TIFS.2017.2788000 10.21437/Interspeech.2020-2248 10.21437/SSW.2019-28 10.1109/MIPRO.2014.6859761 10.1109/ICASSP.2015.7178964 10.21437/Interspeech.2019-2647
ContentType	Journal Article
Copyright	Copyright The Institute of Electrical and Electronics Engineers, Inc. (IEEE) 2022 licence_http://creativecommons.org/publicdomain/zero
Copyright_xml	– notice: Copyright The Institute of Electrical and Electronics Engineers, Inc. (IEEE) 2022 – notice: licence_http://creativecommons.org/publicdomain/zero
DBID	97E RIA RIE AAYXX CITATION 7SC 7T9 8FD JQ2 L7M L~C L~D 1XC VOOES
DOI	10.1109/TASLP.2022.3190741
DatabaseName	IEEE All-Society Periodicals Package (ASPP) 2005–Present IEEE All-Society Periodicals Package (ASPP) 1998–Present IEEE Electronic Library (IEL) CrossRef Computer and Information Systems Abstracts Linguistics and Language Behavior Abstracts (LLBA) Technology Research Database ProQuest Computer Science Collection Advanced Technologies Database with Aerospace Computer and Information Systems Abstracts Academic Computer and Information Systems Abstracts Professional Hyper Article en Ligne (HAL) Hyper Article en Ligne (HAL) (Open Access)
DatabaseTitle	CrossRef Technology Research Database Computer and Information Systems Abstracts – Academic ProQuest Computer Science Collection Computer and Information Systems Abstracts Linguistics and Language Behavior Abstracts (LLBA) Advanced Technologies Database with Aerospace Computer and Information Systems Abstracts Professional
DatabaseTitleList	Technology Research Database
Database_xml	– sequence: 1 dbid: RIE name: IEEE Electronic Library (IEL) url: https://ieeexplore.ieee.org/ sourceTypes: Publisher
DeliveryMethod	fulltext_linktorsrc
Discipline	Engineering Computer Science
EISSN	2329-9304
EndPage	2395
ExternalDocumentID	oai:HAL:hal-03197376v3 10_1109_TASLP_2022_3190741 9829284
Genre	orig-research
GroupedDBID	0R~ 4.4 6IK 97E AAJGR AAKMM AALFJ AARMG AASAJ AAWTH AAWTV ABAZT ABQJQ ABVLG ACIWK ACM ADBCU AEBYY AEFXT AEJOY AENSD AFWIH AFWXC AGQYO AGSQL AHBIQ AIKLT AKJIK AKQYR AKRVB ALMA_UNASSIGNED_HOLDINGS BEFXN BFFAM BGNUA BKEBE BPEOZ CCLIF EBS EJD GUFHI HGAVV IFIPE IPLJI JAVBF LHSKQ M43 OCL PQQKQ RIA RIE RNS ROL AAYXX CITATION 7SC 7T9 8FD JQ2 L7M L~C L~D 1XC VOOES
ID	FETCH-LOGICAL-c439t-5d12fe05e7b665cc81c75bb55d8b40c9bb01fbefaafbd3312e3251b98b04560d3
IEDL.DBID	RIE
ISICitedReferencesCount	22
ISICitedReferencesURI	http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=000833754800004&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
ISSN	2329-9290
IngestDate	Wed Nov 05 07:44:08 EST 2025 Mon Jun 30 04:05:46 EDT 2025 Sat Nov 29 02:44:47 EST 2025 Tue Nov 18 22:38:30 EST 2025 Wed Aug 27 02:25:27 EDT 2025
IsDoiOpenAccess	true
IsOpenAccess	true
IsPeerReviewed	true
IsScholarly	true
Keywords	linkability privacy utility speech recognition speaker identification speaker anonymization
Language	English
License	https://ieeexplore.ieee.org/Xplorehelp/downloads/license-information/IEEE.html https://doi.org/10.15223/policy-029 https://doi.org/10.15223/policy-037 licence_http://creativecommons.org/publicdomain/zero/: http://creativecommons.org/publicdomain/zero
LinkModel	DirectLink
MergedId	FETCHMERGED-LOGICAL-c439t-5d12fe05e7b665cc81c75bb55d8b40c9bb01fbefaafbd3312e3251b98b04560d3
Notes	ObjectType-Article-1 SourceType-Scholarly Journals-1 ObjectType-Feature-2 content type line 14
ORCID	0000-0001-8246-0606 0000-0003-2752-3955 0000-0002-0624-2903 0000-0001-7531-7123 0000-0002-0183-7289 0000-0003-2838-4408 0000-0003-3440-1251 0000-0001-6473-7173
OpenAccessLink	https://inria.hal.science/hal-03197376
PQID	2696284646
PQPubID	85426
PageCount	13
ParticipantIDs	proquest_journals_2696284646 crossref_primary_10_1109_TASLP_2022_3190741 crossref_citationtrail_10_1109_TASLP_2022_3190741 hal_primary_oai_HAL_hal_03197376v3 ieee_primary_9829284
PublicationCentury	2000
PublicationDate	2022-01-01
PublicationDateYYYYMMDD	2022-01-01
PublicationDate_xml	– month: 01 year: 2022 text: 2022-01-01 day: 01
PublicationDecade	2020
PublicationPlace	Piscataway
PublicationPlace_xml	– name: Piscataway
PublicationTitle	IEEE/ACM transactions on audio, speech, and language processing
PublicationTitleAbbrev	TASLP
PublicationYear	2022
Publisher	IEEE The Institute of Electrical and Electronics Engineers, Inc. (IEEE) Institute of Electrical and Electronics Engineers
Publisher_xml	– name: IEEE – name: The Institute of Electrical and Electronics Engineers, Inc. (IEEE) – name: Institute of Electrical and Electronics Engineers
References	ref13 ref12 ref15 ref14 ref52 ref11 ref10 Ardila (ref37) 2020 ref17 ref16 ref19 ref18 ref51 ref50 ref45 ref48 ref47 ref42 ref41 Qian (ref8) 2017 ref49 ref7 Abdi (ref43) 2010 ref9 ref4 ref3 Srivastava (ref38) 2021 ref6 ref5 ref40 Tomashenko (ref22) 2020 Turner (ref25) 2020 ref35 ref34 ref36 ref30 ref33 ref32 ref2 Kenny (ref28) 2010 ref1 ref39 ref24 ref23 ref26 ref20 Farrs (ref44) 2008 ref21 Champion (ref46) 2021 ref27 ref29 Dueck (ref31) 2009
References_xml	– ident: ref9 doi: 10.1109/ASRU.2009.5373356 – ident: ref23 doi: 10.21437/Interspeech.2020-1887 – ident: ref35 doi: 10.21437/Interspeech.2018-1929 – ident: ref3 doi: 10.1109/ICASSP.2016.7472923 – ident: ref15 doi: 10.1109/ICME46284.2020.9102875 – ident: ref26 doi: 10.1016/j.csl.2022.101362 – ident: ref12 doi: 10.1109/FG.2015.7285021 – ident: ref27 doi: 10.1007/11744085_41 – year: 2020 ident: ref25 article-title: Speaker anonymization with distribution-preserving x-vector generation for the VoicePrivacy challenge 2020 – ident: ref16 doi: 10.1109/ICASSP40776.2020.9053868 – ident: ref52 doi: 10.1561/9781601988195 – ident: ref1 doi: 10.1016/0167-6393(95)00009-D – ident: ref30 doi: 10.1109/ICASSP.2014.6853882 – ident: ref17 doi: 10.21437/Interspeech.2020-2692 – year: 2017 ident: ref8 article-title: VoiceMask: Anonymize and sanitize voice input on mobile devices – ident: ref29 doi: 10.21437/Odyssey.2016-59 – ident: ref32 doi: 10.1109/ASRU46091.2019.9003979 – start-page: 1 volume-title: Proc. Encyclopedia Res. Des. year: 2010 ident: ref43 article-title: Jackknife – ident: ref4 doi: 10.1016/j.csl.2019.05.005 – ident: ref47 doi: 10.1007/978-3-540-71050-9 – ident: ref18 doi: 10.1016/j.csl.2017.05.001 – ident: ref7 doi: 10.1109/ICASSP.2016.7472729 – start-page: 4218 volume-title: Proc. 12th Conf. Lang. Resour. Eval. year: 2020 ident: ref37 article-title: Common voice: A massively-multilingual speech corpus – ident: ref19 doi: 10.1109/ICASSP.2018.8461375 – ident: ref39 doi: 10.21437/Interspeech.2008-644 – ident: ref49 doi: 10.1016/j.csl.2019.101024 – ident: ref36 doi: 10.21437/Interspeech.2019-2441 – volume-title: Proc. Odyssey year: 2008 ident: ref44 article-title: How vulnerable are prosodic features to professional imitators? – ident: ref14 doi: 10.21437/Interspeech.2019-2415 – ident: ref6 doi: 10.21437/Interspeech.2020-1333 – ident: ref45 doi: 10.1109/FSKD.2007.347 – ident: ref11 doi: 10.21437/Odyssey.2018-36 – ident: ref21 doi: 10.21437/Interspeech.2011-91 – ident: ref2 doi: 10.1109/ICASSP39728.2021.9413386 – ident: ref42 doi: 10.1016/j.csl.2005.08.001 – volume-title: Proc. Odyssey year: 2010 ident: ref28 article-title: Bayesian speaker verification with heavy-tailed priors – volume-title: Proc. 2nd AAAI Workshop Privacy-Preserving Artif. Intell. year: 2021 ident: ref46 article-title: A study of F0 modification for x-vector based speech pseudonymization across gender – ident: ref20 doi: 10.1109/ICASSP.2019.8682298 – ident: ref48 doi: 10.21437/Interspeech.2021-1070 – ident: ref51 doi: 10.21437/Interspeech.2019-1572 – ident: ref34 doi: 10.21437/Interspeech.2017-950 – year: 2020 ident: ref22 article-title: The VoicePrivacy 2020 challenge evaluation plan – ident: ref50 doi: 10.21437/Interspeech.2020-1090 – ident: ref24 doi: 10.1016/j.csl.2022.101351 – ident: ref40 doi: 10.1109/TIFS.2017.2788000 – ident: ref41 doi: 10.21437/Interspeech.2020-2248 – year: 2021 ident: ref38 article-title: Speaker anonymization Representation, evaluation and formal guarantees – ident: ref13 doi: 10.21437/SSW.2019-28 – ident: ref10 doi: 10.1109/MIPRO.2014.6859761 – year: 2009 ident: ref31 article-title: Affinity propagation: Clustering data by passing messages – ident: ref33 doi: 10.1109/ICASSP.2015.7178964 – ident: ref5 doi: 10.21437/Interspeech.2019-2647
SSID	ssj0001079974
Score	2.402441
Snippet	We study the scenario where individuals ( speakers ) contribute to the publication of an anonymized speech corpus. Data users leverage this public corpus for... We study the scenario where individuals (speakers) contribute to the publication of an anonymized speech corpus. Data users then leverage this public corpus to...
SourceID	hal proquest crossref ieee
SourceType	Open Access Repository Aggregation Database Enrichment Source Index Database Publisher
StartPage	2383
SubjectTerms	Artificial Intelligence Audio equipment Automatic speech recognition Cloning Computation and Language Computer Science Corpus linguistics Error analysis Feature extraction Linkability Machine Learning Measurement Privacy Quality assessment speaker anonymization Speech recognition Speech synthesis Task analysis Training voice conversion Voice recognition
Title	Privacy and Utility of X-Vector Based Speaker Anonymization
URI	https://ieeexplore.ieee.org/document/9829284 https://www.proquest.com/docview/2696284646 https://inria.hal.science/hal-03197376
Volume	30
WOSCitedRecordID	wos000833754800004&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
hasFullText	1
inHoldings	1
isFullTextHit
isPrint
journalDatabaseRights	– providerCode: PRVIEE databaseName: IEEE Electronic Library (IEL) customDbUrl: eissn: 2329-9304 dateEnd: 99991231 omitProxy: false ssIdentifier: ssj0001079974 issn: 2329-9290 databaseCode: RIE dateStart: 20140101 isFulltext: true titleUrlDefault: https://ieeexplore.ieee.org/ providerName: IEEE
link	http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlR3ZSiQxsHDEB33wFseLIL6tWdNX0sGnURQfBhnwYN6aHNWs7KIyjgP-vZV0z6yysuBb06nQoSp1dh0AR1bZjJaQG59nPK-N4lZI5JqUU-GkJackTi3pq-vrcjjUgzk4ntXCIGJMPsOf4TH-y_dP7jWEyk50mWoSpx3oKKWaWq2_8RShtI5Nl8lG0Jy0vpjWyAh9ctu76Q_IG0xTclKDP5h80kOdXyELMo5X-UcmR0VzufK9I67CcmtQsl5zA9ZgDh_XYelDm8ENOB2MHibGvTHz6NndOGTDvrGnmg35fYzZszNSZZ7dPKP5jSPWRATa-sxNuLu8uD2_4u3QBO7IthjzwidpjaJAZaUsnCsTpwpri8KXNhdOWyuS2mJtTG19liUpZmTiWF3aYNwJn23BPH0Gt4EZqxCJ6X0mkLzonHhVpqkjIyGgW5guJFMUVq7tKB4GW_ypomchdBXRXgW0Vy3au_Bjtue56afxX-hDoswMMLTCvur1q_AuVF8pko6TrAsbgQ4zqJYEXdibErJq-fKlSqWWtCZzufP1rl1YDAdogix7MD8eveI-LLjJ-OFldBCv3DswF9Cd
linkProvider	IEEE
linkToHtml	http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlR1da9RAcGirYH3wo7V4bdVFfNO1m02yyeLTKZYT43HQq9zbsh8TLEpbrteD_ntnN7lTsQi-hewsWWZ2PjMfAK9c5XJaQm5DkfOitRV3QiHXpJxKrxw5JWlqSVONx_Vspicb8GZdC4OIKfkM38bH9C8_XPjrGCo70rXUJE434U5ZFDLrqrV-RVREpXVqu0xWguak98WqSkboo-nwpJmQPygluanRI8z-0ESb32IeZBqw8pdUTqrm-OH_HfIRPOhNSjbs7sBj2MDzHbj_W6PBXXg3mZ8trb9h9jyw00XMh71hFy2b8a8pas_ekzIL7OQS7Xecsy4m0FdoPoHT44_TDyPej03gnqyLBS9DJlsUJVZOqdL7OvNV6VxZhtoVwmvnRNY6bK1tXcjzTGJORo7TtYvmnQj5HmzRZ_ApMOsqRGL7kAskP7ogblVSejITIrqFHUC2QqHxfU_xONrih0m-hdAmod1EtJse7QN4vd5z2XXU-Cf0S6LMGjA2wx4NGxPfxfqriuTjMh_AbqTDGqonwQAOV4Q0PWdeGam0ojVVqP3bd72Ae6Ppl8Y0n8afD2A7HqYLuRzC1mJ-jc_grl8uzq7mz9P1-wk3M9Pk
openUrl	ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&rfr_id=info%3Asid%2Fsummon.serialssolutions.com&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Ajournal&rft.genre=article&rft.atitle=Privacy+and+Utility+of+X-Vector+Based+Speaker+Anonymization&rft.jtitle=IEEE%2FACM+transactions+on+audio%2C+speech%2C+and+language+processing&rft.au=Srivastava%2C+Brij+Mohan+Lal&rft.au=Maouche%2C+Mohamed&rft.au=Sahidullah%2C+Md&rft.au=Vincent%2C+Emmanuel&rft.date=2022-01-01&rft.issn=2329-9290&rft.eissn=2329-9304&rft.volume=30&rft.spage=2383&rft.epage=2395&rft_id=info:doi/10.1109%2FTASLP.2022.3190741&rft.externalDBID=n%2Fa&rft.externalDocID=10_1109_TASLP_2022_3190741
thumbnail_l	http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=2329-9290&client=summon
thumbnail_m	http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=2329-9290&client=summon
thumbnail_s	http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=2329-9290&client=summon