Use of Paraconsistent Feature Engineering to support the Long Term Feature choice for Speaker Verification
Autor(a) principal: | |
---|---|
Data de Publicação: | 2021 |
Outros Autores: | , |
Tipo de documento: | Artigo de conferência |
Idioma: | eng |
Título da fonte: | Repositório Institucional da UNESP |
Texto Completo: | http://dx.doi.org/10.32473/flairs.v34i1.128370 http://hdl.handle.net/11449/239854 |
Resumo: | The state-of-art models for speech synthesis and voice conversion can generate synthetic speech perceptually indistinguishable from human speech, and speaker verification is crucial to prevent breaches. The building feature that best distinguishes genuine speech between spoof attacks is an open research subject. We used the baseline ASVSpoof2017, Transfer Learning (TL) set, and Symlet and Daubechies Discrete Wavelet Packet Transform (DWPT) for this investigation. To qualitatively assess the features, we used Paraconsistent Feature Engineering (PFE). Our experiments pointed out that for the use of more robust classifiers, the best choice would be the AlexNet method, while in terms of classification regarding the Equal Error Rate metric, the best suggestion would be Daubechies filter support 21. Finally, our findings indicate that Symlet filter support 17 as the most promising feature, which is evidence that PFE is a useful tool and contributes to feature selection. |
id |
UNSP_19a9bc3c10470f8d98260047e5e7b352 |
---|---|
oai_identifier_str |
oai:repositorio.unesp.br:11449/239854 |
network_acronym_str |
UNSP |
network_name_str |
Repositório Institucional da UNESP |
repository_id_str |
2946 |
spelling |
Use of Paraconsistent Feature Engineering to support the Long Term Feature choice for Speaker VerificationThe state-of-art models for speech synthesis and voice conversion can generate synthetic speech perceptually indistinguishable from human speech, and speaker verification is crucial to prevent breaches. The building feature that best distinguishes genuine speech between spoof attacks is an open research subject. We used the baseline ASVSpoof2017, Transfer Learning (TL) set, and Symlet and Daubechies Discrete Wavelet Packet Transform (DWPT) for this investigation. To qualitatively assess the features, we used Paraconsistent Feature Engineering (PFE). Our experiments pointed out that for the use of more robust classifiers, the best choice would be the AlexNet method, while in terms of classification regarding the Equal Error Rate metric, the best suggestion would be Daubechies filter support 21. Finally, our findings indicate that Symlet filter support 17 as the most promising feature, which is evidence that PFE is a useful tool and contributes to feature selection.Departament of Computer Science São Paulo State University, Rua Cristóvão Colombo 2265, Jd Nazareth,SPDepartament of Computer Science São Paulo State University, Rua Cristóvão Colombo 2265, Jd Nazareth,SPUniversidade Estadual Paulista (UNESP)Almeida, Alex M. G. de[UNESP]Recco, Claudineia H. [UNESP]Guido, Rodrigo C. [UNESP]2023-03-01T19:50:26Z2023-03-01T19:50:26Z2021-01-01info:eu-repo/semantics/publishedVersioninfo:eu-repo/semantics/conferenceObjecthttp://dx.doi.org/10.32473/flairs.v34i1.128370Proceedings of the International Florida Artificial Intelligence Research Society Conference, FLAIRS, v. 34.2334-07622334-0754http://hdl.handle.net/11449/23985410.32473/flairs.v34i1.1283702-s2.0-85125012436Scopusreponame:Repositório Institucional da UNESPinstname:Universidade Estadual Paulista (UNESP)instacron:UNESPengProceedings of the International Florida Artificial Intelligence Research Society Conference, FLAIRSinfo:eu-repo/semantics/openAccess2023-03-01T19:50:26Zoai:repositorio.unesp.br:11449/239854Repositório InstitucionalPUBhttp://repositorio.unesp.br/oai/requestopendoar:29462024-08-05T23:14:22.853799Repositório Institucional da UNESP - Universidade Estadual Paulista (UNESP)false |
dc.title.none.fl_str_mv |
Use of Paraconsistent Feature Engineering to support the Long Term Feature choice for Speaker Verification |
title |
Use of Paraconsistent Feature Engineering to support the Long Term Feature choice for Speaker Verification |
spellingShingle |
Use of Paraconsistent Feature Engineering to support the Long Term Feature choice for Speaker Verification Almeida, Alex M. G. de[UNESP] |
title_short |
Use of Paraconsistent Feature Engineering to support the Long Term Feature choice for Speaker Verification |
title_full |
Use of Paraconsistent Feature Engineering to support the Long Term Feature choice for Speaker Verification |
title_fullStr |
Use of Paraconsistent Feature Engineering to support the Long Term Feature choice for Speaker Verification |
title_full_unstemmed |
Use of Paraconsistent Feature Engineering to support the Long Term Feature choice for Speaker Verification |
title_sort |
Use of Paraconsistent Feature Engineering to support the Long Term Feature choice for Speaker Verification |
author |
Almeida, Alex M. G. de[UNESP] |
author_facet |
Almeida, Alex M. G. de[UNESP] Recco, Claudineia H. [UNESP] Guido, Rodrigo C. [UNESP] |
author_role |
author |
author2 |
Recco, Claudineia H. [UNESP] Guido, Rodrigo C. [UNESP] |
author2_role |
author author |
dc.contributor.none.fl_str_mv |
Universidade Estadual Paulista (UNESP) |
dc.contributor.author.fl_str_mv |
Almeida, Alex M. G. de[UNESP] Recco, Claudineia H. [UNESP] Guido, Rodrigo C. [UNESP] |
description |
The state-of-art models for speech synthesis and voice conversion can generate synthetic speech perceptually indistinguishable from human speech, and speaker verification is crucial to prevent breaches. The building feature that best distinguishes genuine speech between spoof attacks is an open research subject. We used the baseline ASVSpoof2017, Transfer Learning (TL) set, and Symlet and Daubechies Discrete Wavelet Packet Transform (DWPT) for this investigation. To qualitatively assess the features, we used Paraconsistent Feature Engineering (PFE). Our experiments pointed out that for the use of more robust classifiers, the best choice would be the AlexNet method, while in terms of classification regarding the Equal Error Rate metric, the best suggestion would be Daubechies filter support 21. Finally, our findings indicate that Symlet filter support 17 as the most promising feature, which is evidence that PFE is a useful tool and contributes to feature selection. |
publishDate |
2021 |
dc.date.none.fl_str_mv |
2021-01-01 2023-03-01T19:50:26Z 2023-03-01T19:50:26Z |
dc.type.status.fl_str_mv |
info:eu-repo/semantics/publishedVersion |
dc.type.driver.fl_str_mv |
info:eu-repo/semantics/conferenceObject |
format |
conferenceObject |
status_str |
publishedVersion |
dc.identifier.uri.fl_str_mv |
http://dx.doi.org/10.32473/flairs.v34i1.128370 Proceedings of the International Florida Artificial Intelligence Research Society Conference, FLAIRS, v. 34. 2334-0762 2334-0754 http://hdl.handle.net/11449/239854 10.32473/flairs.v34i1.128370 2-s2.0-85125012436 |
url |
http://dx.doi.org/10.32473/flairs.v34i1.128370 http://hdl.handle.net/11449/239854 |
identifier_str_mv |
Proceedings of the International Florida Artificial Intelligence Research Society Conference, FLAIRS, v. 34. 2334-0762 2334-0754 10.32473/flairs.v34i1.128370 2-s2.0-85125012436 |
dc.language.iso.fl_str_mv |
eng |
language |
eng |
dc.relation.none.fl_str_mv |
Proceedings of the International Florida Artificial Intelligence Research Society Conference, FLAIRS |
dc.rights.driver.fl_str_mv |
info:eu-repo/semantics/openAccess |
eu_rights_str_mv |
openAccess |
dc.source.none.fl_str_mv |
Scopus reponame:Repositório Institucional da UNESP instname:Universidade Estadual Paulista (UNESP) instacron:UNESP |
instname_str |
Universidade Estadual Paulista (UNESP) |
instacron_str |
UNESP |
institution |
UNESP |
reponame_str |
Repositório Institucional da UNESP |
collection |
Repositório Institucional da UNESP |
repository.name.fl_str_mv |
Repositório Institucional da UNESP - Universidade Estadual Paulista (UNESP) |
repository.mail.fl_str_mv |
|
_version_ |
1808129501847093248 |