The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. ex. Some numerals are expressed as "XNUMX".
Copyrights notice
The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. Copyrights notice
Nesta carta, propomos uma nova abordagem para o reconhecimento de falantes baseado em SVM, que utiliza um tipo de nova informação fonotática como recurso para modelagem SVM. Os modelos de mistura gaussiana (GMMs) têm se mostrado extremamente bem-sucedidos no reconhecimento de falantes independentes de texto. O modelo de fundo universal GMM (UBM) é um modelo independente de locutor, cada componente do qual pode ser considerado como uma modelagem de algumas classes de sons fonéticos subjacentes. Assumimos que os enunciados de diferentes falantes devem obter probabilidades posteriores médias diferentes no mesmo componente gaussiano do UBM, e o supervetor composto pelas probabilidades posteriores médias em todos os componentes do UBM para cada enunciado deve ser discriminativo. Usamos esses supervetores como recursos para reconhecimento de alto-falante baseado em SVM. Os resultados da experiência em uma tarefa do NIST SRE 2006 mostram que a abordagem proposta demonstra desempenho comparável com os sistemas comumente usados. Os resultados da fusão também são apresentados.
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
Copiar
Xiang ZHANG, Hongbin SUO, Qingwei ZHAO, Yonghong YAN, "Using a Kind of Novel Phonotactic Information for SVM Based Speaker Recognition" in IEICE TRANSACTIONS on Information,
vol. E92-D, no. 4, pp. 746-749, April 2009, doi: 10.1587/transinf.E92.D.746.
Abstract: In this letter, we propose a new approach to SVM based speaker recognition, which utilizes a kind of novel phonotactic information as the feature for SVM modeling. Gaussian mixture models (GMMs) have been proven extremely successful for text-independent speaker recognition. The GMM universal background model (UBM) is a speaker-independent model, each component of which can be considered as modeling some underlying phonetic sound classes. We assume that the utterances from different speakers should get different average posterior probabilities on the same Gaussian component of the UBM, and the supervector composed of the average posterior probabilities on all components of the UBM for each utterance should be discriminative. We use these supervectors as the features for SVM based speaker recognition. Experiment results on a NIST SRE 2006 task show that the proposed approach demonstrates comparable performance with the commonly used systems. Fusion results are also presented.
URL: https://global.ieice.org/en_transactions/information/10.1587/transinf.E92.D.746/_p
Copiar
@ARTICLE{e92-d_4_746,
author={Xiang ZHANG, Hongbin SUO, Qingwei ZHAO, Yonghong YAN, },
journal={IEICE TRANSACTIONS on Information},
title={Using a Kind of Novel Phonotactic Information for SVM Based Speaker Recognition},
year={2009},
volume={E92-D},
number={4},
pages={746-749},
abstract={In this letter, we propose a new approach to SVM based speaker recognition, which utilizes a kind of novel phonotactic information as the feature for SVM modeling. Gaussian mixture models (GMMs) have been proven extremely successful for text-independent speaker recognition. The GMM universal background model (UBM) is a speaker-independent model, each component of which can be considered as modeling some underlying phonetic sound classes. We assume that the utterances from different speakers should get different average posterior probabilities on the same Gaussian component of the UBM, and the supervector composed of the average posterior probabilities on all components of the UBM for each utterance should be discriminative. We use these supervectors as the features for SVM based speaker recognition. Experiment results on a NIST SRE 2006 task show that the proposed approach demonstrates comparable performance with the commonly used systems. Fusion results are also presented.},
keywords={},
doi={10.1587/transinf.E92.D.746},
ISSN={1745-1361},
month={April},}
Copiar
TY - JOUR
TI - Using a Kind of Novel Phonotactic Information for SVM Based Speaker Recognition
T2 - IEICE TRANSACTIONS on Information
SP - 746
EP - 749
AU - Xiang ZHANG
AU - Hongbin SUO
AU - Qingwei ZHAO
AU - Yonghong YAN
PY - 2009
DO - 10.1587/transinf.E92.D.746
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E92-D
IS - 4
JA - IEICE TRANSACTIONS on Information
Y1 - April 2009
AB - In this letter, we propose a new approach to SVM based speaker recognition, which utilizes a kind of novel phonotactic information as the feature for SVM modeling. Gaussian mixture models (GMMs) have been proven extremely successful for text-independent speaker recognition. The GMM universal background model (UBM) is a speaker-independent model, each component of which can be considered as modeling some underlying phonetic sound classes. We assume that the utterances from different speakers should get different average posterior probabilities on the same Gaussian component of the UBM, and the supervector composed of the average posterior probabilities on all components of the UBM for each utterance should be discriminative. We use these supervectors as the features for SVM based speaker recognition. Experiment results on a NIST SRE 2006 task show that the proposed approach demonstrates comparable performance with the commonly used systems. Fusion results are also presented.
ER -