Combining standard and throat microphones for robust speech recognition

Martin Graciarena, Horacio Franco, Mustafa (Kemal) Sonmez, Harry Bratt

Research output: Contribution to journalArticle

78 Citations (Scopus)

Abstract

We present a method to combine the standard and throat microphone signals for robust speech recognition in noisy environments. Our approach is to use the probabilistic optimum filter (POF) mapping algorithm to estimate the standard microphone clean-speech feature vectors, used by standard speech recognizers, from both microphones' noisy-speech feature vectors. A small untranscribed "stereo" database (noisy and clean simultaneous recordings) is required to train the POF mappings. In continuous-speech recognition experiments using SRI International's DECIPHER recognition system, both using artificially added noise and using recorded noisy speech, the combined-microphone approach significantly outperforms the single-microphone approach.

Original languageEnglish (US)
Pages (from-to)72-74
Number of pages3
JournalIEEE Signal Processing Letters
Volume10
Issue number3
DOIs
StatePublished - Mar 2003
Externally publishedYes

Fingerprint

Robust Speech Recognition
Microphones
Speech recognition
Feature Vector
Filter
Continuous speech recognition
Speech Recognition
Standards
Speech
Estimate
Experiment
Experiments

Keywords

  • Noise robustness
  • Probabilistic optimum filtering
  • Speech recognition
  • Throat microphone

ASJC Scopus subject areas

  • Signal Processing
  • Electrical and Electronic Engineering

Cite this

Combining standard and throat microphones for robust speech recognition. / Graciarena, Martin; Franco, Horacio; Sonmez, Mustafa (Kemal); Bratt, Harry.

In: IEEE Signal Processing Letters, Vol. 10, No. 3, 03.2003, p. 72-74.

Research output: Contribution to journalArticle

Graciarena, Martin ; Franco, Horacio ; Sonmez, Mustafa (Kemal) ; Bratt, Harry. / Combining standard and throat microphones for robust speech recognition. In: IEEE Signal Processing Letters. 2003 ; Vol. 10, No. 3. pp. 72-74.
@article{6c596fe8fa794bd190e92b3ae785ac72,
title = "Combining standard and throat microphones for robust speech recognition",
abstract = "We present a method to combine the standard and throat microphone signals for robust speech recognition in noisy environments. Our approach is to use the probabilistic optimum filter (POF) mapping algorithm to estimate the standard microphone clean-speech feature vectors, used by standard speech recognizers, from both microphones' noisy-speech feature vectors. A small untranscribed {"}stereo{"} database (noisy and clean simultaneous recordings) is required to train the POF mappings. In continuous-speech recognition experiments using SRI International's DECIPHER recognition system, both using artificially added noise and using recorded noisy speech, the combined-microphone approach significantly outperforms the single-microphone approach.",
keywords = "Noise robustness, Probabilistic optimum filtering, Speech recognition, Throat microphone",
author = "Martin Graciarena and Horacio Franco and Sonmez, {Mustafa (Kemal)} and Harry Bratt",
year = "2003",
month = "3",
doi = "10.1109/LSP.2003.808549",
language = "English (US)",
volume = "10",
pages = "72--74",
journal = "IEEE Signal Processing Letters",
issn = "1070-9908",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
number = "3",

}

TY - JOUR

T1 - Combining standard and throat microphones for robust speech recognition

AU - Graciarena, Martin

AU - Franco, Horacio

AU - Sonmez, Mustafa (Kemal)

AU - Bratt, Harry

PY - 2003/3

Y1 - 2003/3

N2 - We present a method to combine the standard and throat microphone signals for robust speech recognition in noisy environments. Our approach is to use the probabilistic optimum filter (POF) mapping algorithm to estimate the standard microphone clean-speech feature vectors, used by standard speech recognizers, from both microphones' noisy-speech feature vectors. A small untranscribed "stereo" database (noisy and clean simultaneous recordings) is required to train the POF mappings. In continuous-speech recognition experiments using SRI International's DECIPHER recognition system, both using artificially added noise and using recorded noisy speech, the combined-microphone approach significantly outperforms the single-microphone approach.

AB - We present a method to combine the standard and throat microphone signals for robust speech recognition in noisy environments. Our approach is to use the probabilistic optimum filter (POF) mapping algorithm to estimate the standard microphone clean-speech feature vectors, used by standard speech recognizers, from both microphones' noisy-speech feature vectors. A small untranscribed "stereo" database (noisy and clean simultaneous recordings) is required to train the POF mappings. In continuous-speech recognition experiments using SRI International's DECIPHER recognition system, both using artificially added noise and using recorded noisy speech, the combined-microphone approach significantly outperforms the single-microphone approach.

KW - Noise robustness

KW - Probabilistic optimum filtering

KW - Speech recognition

KW - Throat microphone

UR - http://www.scopus.com/inward/record.url?scp=0037341762&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0037341762&partnerID=8YFLogxK

U2 - 10.1109/LSP.2003.808549

DO - 10.1109/LSP.2003.808549

M3 - Article

AN - SCOPUS:0037341762

VL - 10

SP - 72

EP - 74

JO - IEEE Signal Processing Letters

JF - IEEE Signal Processing Letters

SN - 1070-9908

IS - 3

ER -