An analysis of the influence of deep neural network (DNN) topology in bottleneck feature based language recognition.
Language recognition systems based on bottleneck features have recently become the state-of-the-art in this research field, showing its success in the last Language Recognition Evaluation (LRE 2015) organized by NIST (U.S. National Institute of Standards and Technology). This type of system is based...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Public Library of Science (PLoS)
2017-01-01
|
| Series: | PLoS ONE |
| Online Access: | http://europepmc.org/articles/PMC5552160?pdf=render |
| _version_ | 1818559799195860992 |
|---|---|
| author | Alicia Lozano-Diez Ruben Zazo Doroteo T Toledano Joaquin Gonzalez-Rodriguez |
| author_facet | Alicia Lozano-Diez Ruben Zazo Doroteo T Toledano Joaquin Gonzalez-Rodriguez |
| author_sort | Alicia Lozano-Diez |
| collection | DOAJ |
| description | Language recognition systems based on bottleneck features have recently become the state-of-the-art in this research field, showing its success in the last Language Recognition Evaluation (LRE 2015) organized by NIST (U.S. National Institute of Standards and Technology). This type of system is based on a deep neural network (DNN) trained to discriminate between phonetic units, i.e. trained for the task of automatic speech recognition (ASR). This DNN aims to compress information in one of its layers, known as bottleneck (BN) layer, which is used to obtain a new frame representation of the audio signal. This representation has been proven to be useful for the task of language identification (LID). Thus, bottleneck features are used as input to the language recognition system, instead of a classical parameterization of the signal based on cepstral feature vectors such as MFCCs (Mel Frequency Cepstral Coefficients). Despite the success of this approach in language recognition, there is a lack of studies analyzing in a systematic way how the topology of the DNN influences the performance of bottleneck feature-based language recognition systems. In this work, we try to fill-in this gap, analyzing language recognition results with different topologies for the DNN used to extract the bottleneck features, comparing them and against a reference system based on a more classical cepstral representation of the input signal with a total variability model. This way, we obtain useful knowledge about how the DNN configuration influences bottleneck feature-based language recognition systems performance. |
| first_indexed | 2024-12-14T00:30:14Z |
| format | Article |
| id | doaj.art-dc2eea914ea3440786006471faab3e48 |
| institution | Directory Open Access Journal |
| issn | 1932-6203 |
| language | English |
| last_indexed | 2024-12-14T00:30:14Z |
| publishDate | 2017-01-01 |
| publisher | Public Library of Science (PLoS) |
| record_format | Article |
| series | PLoS ONE |
| spelling | doaj.art-dc2eea914ea3440786006471faab3e482022-12-21T23:24:54ZengPublic Library of Science (PLoS)PLoS ONE1932-62032017-01-01128e018258010.1371/journal.pone.0182580An analysis of the influence of deep neural network (DNN) topology in bottleneck feature based language recognition.Alicia Lozano-DiezRuben ZazoDoroteo T ToledanoJoaquin Gonzalez-RodriguezLanguage recognition systems based on bottleneck features have recently become the state-of-the-art in this research field, showing its success in the last Language Recognition Evaluation (LRE 2015) organized by NIST (U.S. National Institute of Standards and Technology). This type of system is based on a deep neural network (DNN) trained to discriminate between phonetic units, i.e. trained for the task of automatic speech recognition (ASR). This DNN aims to compress information in one of its layers, known as bottleneck (BN) layer, which is used to obtain a new frame representation of the audio signal. This representation has been proven to be useful for the task of language identification (LID). Thus, bottleneck features are used as input to the language recognition system, instead of a classical parameterization of the signal based on cepstral feature vectors such as MFCCs (Mel Frequency Cepstral Coefficients). Despite the success of this approach in language recognition, there is a lack of studies analyzing in a systematic way how the topology of the DNN influences the performance of bottleneck feature-based language recognition systems. In this work, we try to fill-in this gap, analyzing language recognition results with different topologies for the DNN used to extract the bottleneck features, comparing them and against a reference system based on a more classical cepstral representation of the input signal with a total variability model. This way, we obtain useful knowledge about how the DNN configuration influences bottleneck feature-based language recognition systems performance.http://europepmc.org/articles/PMC5552160?pdf=render |
| spellingShingle | Alicia Lozano-Diez Ruben Zazo Doroteo T Toledano Joaquin Gonzalez-Rodriguez An analysis of the influence of deep neural network (DNN) topology in bottleneck feature based language recognition. PLoS ONE |
| title | An analysis of the influence of deep neural network (DNN) topology in bottleneck feature based language recognition. |
| title_full | An analysis of the influence of deep neural network (DNN) topology in bottleneck feature based language recognition. |
| title_fullStr | An analysis of the influence of deep neural network (DNN) topology in bottleneck feature based language recognition. |
| title_full_unstemmed | An analysis of the influence of deep neural network (DNN) topology in bottleneck feature based language recognition. |
| title_short | An analysis of the influence of deep neural network (DNN) topology in bottleneck feature based language recognition. |
| title_sort | analysis of the influence of deep neural network dnn topology in bottleneck feature based language recognition |
| url | http://europepmc.org/articles/PMC5552160?pdf=render |
| work_keys_str_mv | AT alicialozanodiez ananalysisoftheinfluenceofdeepneuralnetworkdnntopologyinbottleneckfeaturebasedlanguagerecognition AT rubenzazo ananalysisoftheinfluenceofdeepneuralnetworkdnntopologyinbottleneckfeaturebasedlanguagerecognition AT doroteottoledano ananalysisoftheinfluenceofdeepneuralnetworkdnntopologyinbottleneckfeaturebasedlanguagerecognition AT joaquingonzalezrodriguez ananalysisoftheinfluenceofdeepneuralnetworkdnntopologyinbottleneckfeaturebasedlanguagerecognition AT alicialozanodiez analysisoftheinfluenceofdeepneuralnetworkdnntopologyinbottleneckfeaturebasedlanguagerecognition AT rubenzazo analysisoftheinfluenceofdeepneuralnetworkdnntopologyinbottleneckfeaturebasedlanguagerecognition AT doroteottoledano analysisoftheinfluenceofdeepneuralnetworkdnntopologyinbottleneckfeaturebasedlanguagerecognition AT joaquingonzalezrodriguez analysisoftheinfluenceofdeepneuralnetworkdnntopologyinbottleneckfeaturebasedlanguagerecognition |