Spoken language recognition based on senone posteriors
This paper explores in depth a recently proposed approach to spoken language recognition based on the estimated posteriors for a set of senones representing the phonetic space of one or more languages. A neural network (NN) is trained to estimate the posterior probabilities for the senones at a fram...
Guardado en:
| Autores principales: | , , , , , , , , , |
|---|---|
| Formato: | CONF |
| Materias: | |
| Acceso en línea: | http://hdl.handle.net/20.500.12110/paper_2308457X_v_n_p2150_Ferrer |
| Aporte de: |
| id |
todo:paper_2308457X_v_n_p2150_Ferrer |
|---|---|
| record_format |
dspace |
| spelling |
todo:paper_2308457X_v_n_p2150_Ferrer2023-10-03T16:40:55Z Spoken language recognition based on senone posteriors Ferrer, L. Lei, Y. McLaren, M. Scheffer, N. Chng E.S. Li H. Meng H. Ma B. Xie L. Amazon; Baidu; et al.; Google; Temasek Laboratories at Nanyang Technological University (TL at NTU); WeChat Speech communication Activity detection Dimensionality reduction Feature vectors Language recognition Neural network (nn) Posterior probability Speech data Spoken language recognition Speech recognition This paper explores in depth a recently proposed approach to spoken language recognition based on the estimated posteriors for a set of senones representing the phonetic space of one or more languages. A neural network (NN) is trained to estimate the posterior probabilities for the senones at a frame level. A feature vector is then derived for every sample using these posteriors. The effect of the language used in training the NN and the number of senones are studied. Speech-activity detection (SAD) and dimensionality reduction approaches are also explored and Gaussian and NN backends are compared. Results are presented on heavily degraded speech data. The proposed system is shown to give over 40% relative gain compared to a state-of-the-art language recognition system at sample durations from 3 to 120 seconds. Copyright © 2014 ISCA. CONF info:eu-repo/semantics/openAccess http://creativecommons.org/licenses/by/2.5/ar http://hdl.handle.net/20.500.12110/paper_2308457X_v_n_p2150_Ferrer |
| institution |
Universidad de Buenos Aires |
| institution_str |
I-28 |
| repository_str |
R-134 |
| collection |
Biblioteca Digital - Facultad de Ciencias Exactas y Naturales (UBA) |
| topic |
Speech communication Activity detection Dimensionality reduction Feature vectors Language recognition Neural network (nn) Posterior probability Speech data Spoken language recognition Speech recognition |
| spellingShingle |
Speech communication Activity detection Dimensionality reduction Feature vectors Language recognition Neural network (nn) Posterior probability Speech data Spoken language recognition Speech recognition Ferrer, L. Lei, Y. McLaren, M. Scheffer, N. Chng E.S. Li H. Meng H. Ma B. Xie L. Amazon; Baidu; et al.; Google; Temasek Laboratories at Nanyang Technological University (TL at NTU); WeChat Spoken language recognition based on senone posteriors |
| topic_facet |
Speech communication Activity detection Dimensionality reduction Feature vectors Language recognition Neural network (nn) Posterior probability Speech data Spoken language recognition Speech recognition |
| description |
This paper explores in depth a recently proposed approach to spoken language recognition based on the estimated posteriors for a set of senones representing the phonetic space of one or more languages. A neural network (NN) is trained to estimate the posterior probabilities for the senones at a frame level. A feature vector is then derived for every sample using these posteriors. The effect of the language used in training the NN and the number of senones are studied. Speech-activity detection (SAD) and dimensionality reduction approaches are also explored and Gaussian and NN backends are compared. Results are presented on heavily degraded speech data. The proposed system is shown to give over 40% relative gain compared to a state-of-the-art language recognition system at sample durations from 3 to 120 seconds. Copyright © 2014 ISCA. |
| format |
CONF |
| author |
Ferrer, L. Lei, Y. McLaren, M. Scheffer, N. Chng E.S. Li H. Meng H. Ma B. Xie L. Amazon; Baidu; et al.; Google; Temasek Laboratories at Nanyang Technological University (TL at NTU); WeChat |
| author_facet |
Ferrer, L. Lei, Y. McLaren, M. Scheffer, N. Chng E.S. Li H. Meng H. Ma B. Xie L. Amazon; Baidu; et al.; Google; Temasek Laboratories at Nanyang Technological University (TL at NTU); WeChat |
| author_sort |
Ferrer, L. |
| title |
Spoken language recognition based on senone posteriors |
| title_short |
Spoken language recognition based on senone posteriors |
| title_full |
Spoken language recognition based on senone posteriors |
| title_fullStr |
Spoken language recognition based on senone posteriors |
| title_full_unstemmed |
Spoken language recognition based on senone posteriors |
| title_sort |
spoken language recognition based on senone posteriors |
| url |
http://hdl.handle.net/20.500.12110/paper_2308457X_v_n_p2150_Ferrer |
| work_keys_str_mv |
AT ferrerl spokenlanguagerecognitionbasedonsenoneposteriors AT leiy spokenlanguagerecognitionbasedonsenoneposteriors AT mclarenm spokenlanguagerecognitionbasedonsenoneposteriors AT scheffern spokenlanguagerecognitionbasedonsenoneposteriors AT chnges spokenlanguagerecognitionbasedonsenoneposteriors AT lih spokenlanguagerecognitionbasedonsenoneposteriors AT mengh spokenlanguagerecognitionbasedonsenoneposteriors AT mab spokenlanguagerecognitionbasedonsenoneposteriors AT xiel spokenlanguagerecognitionbasedonsenoneposteriors AT amazonbaiduetalgoogletemaseklaboratoriesatnanyangtechnologicaluniversitytlatntuwechat spokenlanguagerecognitionbasedonsenoneposteriors |
| _version_ |
1807321820347498496 |