Noisy Speech Recognition Based on Integration/Selection of Multiple Noise Suppression Methods Using Noise GMMs
Summary :
To achieve high recognition performance for a wide variety of noise and for a wide range of signal-to-noise ratio, this paper presents methods for integration of four noise reduction algorithms: spectral subtraction with smoothing of time direction, temporal domain SVD-based speech enhancement, GMM-based speech estimation and KLT-based comb-filtering. In this paper, we proposed two types of combination methods of noise suppression algorithms: selection of front-end processor and combination of results from multiple recognition processes. Recognition results on the CENSREC-1 task showed the effectiveness of our proposed methods.
- Publication
- IEICE TRANSACTIONS on Information Vol.E91-D No.3 pp.411-421
- Publication Date
- 2008/03/01
- Publicized
- Online ISSN
- 1745-1361
- DOI
- 10.1093/ietisy/e91-d.3.411
- Type of Manuscript
- Special Section PAPER (Special Section on Robust Speech Processing in Realistic Environments)
- Category
- Noisy Speech Recognition
Authors
Keyword
Latest Issue
Copyrights notice of machine-translated contents
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
Cite this
Copy
Norihide KITAOKA, Souta HAMAGUCHI, Seiichi NAKAGAWA, "Noisy Speech Recognition Based on Integration/Selection of Multiple Noise Suppression Methods Using Noise GMMs" in IEICE TRANSACTIONS on Information,
vol. E91-D, no. 3, pp. 411-421, March 2008, doi: 10.1093/ietisy/e91-d.3.411.
Abstract: To achieve high recognition performance for a wide variety of noise and for a wide range of signal-to-noise ratio, this paper presents methods for integration of four noise reduction algorithms: spectral subtraction with smoothing of time direction, temporal domain SVD-based speech enhancement, GMM-based speech estimation and KLT-based comb-filtering. In this paper, we proposed two types of combination methods of noise suppression algorithms: selection of front-end processor and combination of results from multiple recognition processes. Recognition results on the CENSREC-1 task showed the effectiveness of our proposed methods.
URL: https://global.ieice.org/en_transactions/information/10.1093/ietisy/e91-d.3.411/_p
Copy
@ARTICLE{e91-d_3_411,
author={Norihide KITAOKA, Souta HAMAGUCHI, Seiichi NAKAGAWA, },
journal={IEICE TRANSACTIONS on Information},
title={Noisy Speech Recognition Based on Integration/Selection of Multiple Noise Suppression Methods Using Noise GMMs},
year={2008},
volume={E91-D},
number={3},
pages={411-421},
abstract={To achieve high recognition performance for a wide variety of noise and for a wide range of signal-to-noise ratio, this paper presents methods for integration of four noise reduction algorithms: spectral subtraction with smoothing of time direction, temporal domain SVD-based speech enhancement, GMM-based speech estimation and KLT-based comb-filtering. In this paper, we proposed two types of combination methods of noise suppression algorithms: selection of front-end processor and combination of results from multiple recognition processes. Recognition results on the CENSREC-1 task showed the effectiveness of our proposed methods.},
keywords={},
doi={10.1093/ietisy/e91-d.3.411},
ISSN={1745-1361},
month={March},}
Copy
TY - JOUR
TI - Noisy Speech Recognition Based on Integration/Selection of Multiple Noise Suppression Methods Using Noise GMMs
T2 - IEICE TRANSACTIONS on Information
SP - 411
EP - 421
AU - Norihide KITAOKA
AU - Souta HAMAGUCHI
AU - Seiichi NAKAGAWA
PY - 2008
DO - 10.1093/ietisy/e91-d.3.411
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E91-D
IS - 3
JA - IEICE TRANSACTIONS on Information
Y1 - March 2008
AB - To achieve high recognition performance for a wide variety of noise and for a wide range of signal-to-noise ratio, this paper presents methods for integration of four noise reduction algorithms: spectral subtraction with smoothing of time direction, temporal domain SVD-based speech enhancement, GMM-based speech estimation and KLT-based comb-filtering. In this paper, we proposed two types of combination methods of noise suppression algorithms: selection of front-end processor and combination of results from multiple recognition processes. Recognition results on the CENSREC-1 task showed the effectiveness of our proposed methods.
ER -
English
