Publication detail

Autoencoder based multi-stream combination for noise robust speech recognition

MALLIDI, S. OGAWA, T. VESELÝ, K. NIDADAVOLU, P. HEŘMANSKÝ, H.

Original Title

Autoencoder based multi-stream combination for noise robust speech recognition

Type

conference paper

Language

English

Original Abstract

In this paper, we have proposed techniques to estimate performance of DNN based classifiers. The technique is based on modeling the distribution of DNN outputs, using autoencoders.

Keywords

speech recognition, human-computer interaction, computational paralinguistics

Authors

MALLIDI, S.; OGAWA, T.; VESELÝ, K.; NIDADAVOLU, P.; HEŘMANSKÝ, H.

RIV year

2015

Released

6. 9. 2015

Publisher

International Speech Communication Association

Location

Dresden

ISBN

978-1-5108-1790-6

Book

Proceeding of Interspeech 2015

ISBN

1990-9772

Periodical

Proceedings of Interspeech

Year of study

2015

Number

09

State

French Republic

Pages from

3551

Pages to

3555

Pages count

5

URL

http://www.fit.vutbr.cz/research/groups/speech/publi/2015/mallidi_interspeech2015_IS150897.pdf

BibTex

@inproceedings{BUT119906,
  author="Sri Harish {Mallidi} and Tetsuji {Ogawa} and Karel {Veselý} and Phani {Nidadavolu} and Hynek {Heřmanský}",
  title="Autoencoder based multi-stream combination for noise robust speech recognition",
  booktitle="Proceeding of Interspeech 2015",
  year="2015",
  journal="Proceedings of Interspeech",
  volume="2015",
  number="09",
  pages="3551--3555",
  publisher="International Speech Communication Association",
  address="Dresden",
  isbn="978-1-5108-1790-6",
  issn="1990-9772",
  url="http://www.fit.vutbr.cz/research/groups/speech/publi/2015/mallidi_interspeech2015_IS150897.pdf"
}