Recognition of convolutive speech mixtures by missing feature techniques for ICA

Dorothea Kolossa; Hiroshi Sawada; Ramon Fernandez Astudillo; Reinhold Orglmeister; Shoji Makino

doi:10.1109/ACSSC.2006.354987

Recognition of convolutive speech mixtures by missing feature techniques for ICA

Dorothea Kolossa^*, Hiroshi Sawada, Ramon Fernandez Astudillo, Reinhold Orglmeister, Shoji Makino

^*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

9 Citations (Scopus)

Abstract

One challenging problem for robust speech recognition is the cocktail party effect, where multiple speaker signals are active simultaneously in an overlapping frequency range. In that case, independent component analysis (ICA) can separate the signals in reverberant environments, also. However, incurred feature distortions prove detrimental for speech recognition. To reduce consequential recognition errors, we describe the use of ICA for the additional estimation of uncertainty information. This information is subsequently used in missing feature speech recognition, which leads to far more correct and accurate recognition also in reverberant situations at RT₆₀ = 300ms.

Original language	English
Title of host publication	Conference Record of the 40th Asilomar Conference on Signals, Systems and Computers, ACSSC '06
Pages	1397-1401
Number of pages	5
DOIs	https://doi.org/10.1109/ACSSC.2006.354987
Publication status	Published - 2006
Externally published	Yes
Event	40th Asilomar Conference on Signals, Systems, and Computers, ACSSC '06 - Pacific Grove, CA, United States Duration: 2006 Oct 29 → 2006 Nov 1

Publication series

Name	Conference Record - Asilomar Conference on Signals, Systems and Computers
ISSN (Print)	1058-6393

Conference

Conference	40th Asilomar Conference on Signals, Systems, and Computers, ACSSC '06
Country/Territory	United States
City	Pacific Grove, CA
Period	06/10/29 → 06/11/1

ASJC Scopus subject areas

Signal Processing
Computer Networks and Communications

Access to Document

10.1109/ACSSC.2006.354987

Cite this

Kolossa, D., Sawada, H., Astudillo, R. F., Orglmeister, R., & Makino, S. (2006). Recognition of convolutive speech mixtures by missing feature techniques for ICA. In Conference Record of the 40th Asilomar Conference on Signals, Systems and Computers, ACSSC '06 (pp. 1397-1401). Article 4176797 (Conference Record - Asilomar Conference on Signals, Systems and Computers). https://doi.org/10.1109/ACSSC.2006.354987

Recognition of convolutive speech mixtures by missing feature techniques for ICA. / Kolossa, Dorothea; Sawada, Hiroshi; Astudillo, Ramon Fernandez et al.
Conference Record of the 40th Asilomar Conference on Signals, Systems and Computers, ACSSC '06. 2006. p. 1397-1401 4176797 (Conference Record - Asilomar Conference on Signals, Systems and Computers).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Kolossa, D, Sawada, H, Astudillo, RF, Orglmeister, R & Makino, S 2006, Recognition of convolutive speech mixtures by missing feature techniques for ICA. in Conference Record of the 40th Asilomar Conference on Signals, Systems and Computers, ACSSC '06., 4176797, Conference Record - Asilomar Conference on Signals, Systems and Computers, pp. 1397-1401, 40th Asilomar Conference on Signals, Systems, and Computers, ACSSC '06, Pacific Grove, CA, United States, 06/10/29. https://doi.org/10.1109/ACSSC.2006.354987

Kolossa D, Sawada H, Astudillo RF, Orglmeister R, Makino S. Recognition of convolutive speech mixtures by missing feature techniques for ICA. In Conference Record of the 40th Asilomar Conference on Signals, Systems and Computers, ACSSC '06. 2006. p. 1397-1401. 4176797. (Conference Record - Asilomar Conference on Signals, Systems and Computers). doi: 10.1109/ACSSC.2006.354987

@inproceedings{ae0f9c3cfffa46d3b159cb2dad4e35e2,

title = "Recognition of convolutive speech mixtures by missing feature techniques for ICA",

abstract = "One challenging problem for robust speech recognition is the cocktail party effect, where multiple speaker signals are active simultaneously in an overlapping frequency range. In that case, independent component analysis (ICA) can separate the signals in reverberant environments, also. However, incurred feature distortions prove detrimental for speech recognition. To reduce consequential recognition errors, we describe the use of ICA for the additional estimation of uncertainty information. This information is subsequently used in missing feature speech recognition, which leads to far more correct and accurate recognition also in reverberant situations at RT60 = 300ms.",

author = "Dorothea Kolossa and Hiroshi Sawada and Astudillo, {Ramon Fernandez} and Reinhold Orglmeister and Shoji Makino",

year = "2006",

doi = "10.1109/ACSSC.2006.354987",

language = "English",

isbn = "1424407850",

series = "Conference Record - Asilomar Conference on Signals, Systems and Computers",

pages = "1397--1401",

booktitle = "Conference Record of the 40th Asilomar Conference on Signals, Systems and Computers, ACSSC '06",

note = "40th Asilomar Conference on Signals, Systems, and Computers, ACSSC '06 ; Conference date: 29-10-2006 Through 01-11-2006",

}

TY - GEN

T1 - Recognition of convolutive speech mixtures by missing feature techniques for ICA

AU - Kolossa, Dorothea

AU - Sawada, Hiroshi

AU - Astudillo, Ramon Fernandez

AU - Orglmeister, Reinhold

AU - Makino, Shoji

PY - 2006

Y1 - 2006

N2 - One challenging problem for robust speech recognition is the cocktail party effect, where multiple speaker signals are active simultaneously in an overlapping frequency range. In that case, independent component analysis (ICA) can separate the signals in reverberant environments, also. However, incurred feature distortions prove detrimental for speech recognition. To reduce consequential recognition errors, we describe the use of ICA for the additional estimation of uncertainty information. This information is subsequently used in missing feature speech recognition, which leads to far more correct and accurate recognition also in reverberant situations at RT60 = 300ms.

AB - One challenging problem for robust speech recognition is the cocktail party effect, where multiple speaker signals are active simultaneously in an overlapping frequency range. In that case, independent component analysis (ICA) can separate the signals in reverberant environments, also. However, incurred feature distortions prove detrimental for speech recognition. To reduce consequential recognition errors, we describe the use of ICA for the additional estimation of uncertainty information. This information is subsequently used in missing feature speech recognition, which leads to far more correct and accurate recognition also in reverberant situations at RT60 = 300ms.

UR - http://www.scopus.com/inward/record.url?scp=47049124615&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=47049124615&partnerID=8YFLogxK

U2 - 10.1109/ACSSC.2006.354987

DO - 10.1109/ACSSC.2006.354987

M3 - Conference contribution

AN - SCOPUS:47049124615

SN - 1424407850

SN - 9781424407859

T3 - Conference Record - Asilomar Conference on Signals, Systems and Computers

SP - 1397

EP - 1401

BT - Conference Record of the 40th Asilomar Conference on Signals, Systems and Computers, ACSSC '06

T2 - 40th Asilomar Conference on Signals, Systems, and Computers, ACSSC '06

Y2 - 29 October 2006 through 1 November 2006

ER -

Recognition of convolutive speech mixtures by missing feature techniques for ICA

Abstract

Publication series

Conference

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this