Feature Extraction with Combination of HMT-Based Denoising and Weighted Filter Bank Analysis for Robust Speech Recognition

Sungyun JUNG; Jongmok SON; Keunsung BAE

doi:10.1093/ietisy/e88-d.3.435

Feature Extraction with Combination of HMT-Based Denoising and Weighted Filter Bank Analysis for Robust Speech Recognition

Sungyun JUNG, Jongmok SON, Keunsung BAE

Full Text Views

0

Share
Cite this

Summary :

In this paper, we propose a new feature extraction method that combines both HMT-based denoising and weighted filter bank analysis for robust speech recognition. The proposed method is made up of two stages in cascade. The first stage is denoising process based on the wavelet domain Hidden Markov Tree model, and the second one is the filter bank analysis with weighting coefficients obtained from the residual noise in the first stage. To evaluate performance of the proposed method, recognition experiments were carried out for additive white Gaussian and pink noise with signal-to-noise ratio from 25 dB to 0 dB. Experiment results demonstrate the superiority of the proposed method to the conventional ones.

Publication: IEICE TRANSACTIONS on Information Vol.E88-D No.3 pp.435-438

Publication Date: 2005/03/01

Publicized

Online ISSN

DOI: 10.1093/ietisy/e88-d.3.435

Type of Manuscript: Special Section LETTER (Special Section on Corpus-Based Speech Technologies)

Category

Cite this

Copy

Sungyun JUNG, Jongmok SON, Keunsung BAE, "Feature Extraction with Combination of HMT-Based Denoising and Weighted Filter Bank Analysis for Robust Speech Recognition" in IEICE TRANSACTIONS on Information, vol. E88-D, no. 3, pp. 435-438, March 2005, doi: 10.1093/ietisy/e88-d.3.435.
Abstract: In this paper, we propose a new feature extraction method that combines both HMT-based denoising and weighted filter bank analysis for robust speech recognition. The proposed method is made up of two stages in cascade. The first stage is denoising process based on the wavelet domain Hidden Markov Tree model, and the second one is the filter bank analysis with weighting coefficients obtained from the residual noise in the first stage. To evaluate performance of the proposed method, recognition experiments were carried out for additive white Gaussian and pink noise with signal-to-noise ratio from 25 dB to 0 dB. Experiment results demonstrate the superiority of the proposed method to the conventional ones.
URL: https://globals.ieice.org/en_transactions/information/10.1093/ietisy/e88-d.3.435/_p

Copy

@ARTICLE{e88-d_3_435,
author={Sungyun JUNG, Jongmok SON, Keunsung BAE, },
journal={IEICE TRANSACTIONS on Information},
title={Feature Extraction with Combination of HMT-Based Denoising and Weighted Filter Bank Analysis for Robust Speech Recognition},
year={2005},
volume={E88-D},
number={3},
pages={435-438},
abstract={In this paper, we propose a new feature extraction method that combines both HMT-based denoising and weighted filter bank analysis for robust speech recognition. The proposed method is made up of two stages in cascade. The first stage is denoising process based on the wavelet domain Hidden Markov Tree model, and the second one is the filter bank analysis with weighting coefficients obtained from the residual noise in the first stage. To evaluate performance of the proposed method, recognition experiments were carried out for additive white Gaussian and pink noise with signal-to-noise ratio from 25 dB to 0 dB. Experiment results demonstrate the superiority of the proposed method to the conventional ones.},
keywords={},
doi={10.1093/ietisy/e88-d.3.435},
ISSN={},
month={March},}

Copy

TY - JOUR
TI - Feature Extraction with Combination of HMT-Based Denoising and Weighted Filter Bank Analysis for Robust Speech Recognition
T2 - IEICE TRANSACTIONS on Information
SP - 435
EP - 438
AU - Sungyun JUNG
AU - Jongmok SON
AU - Keunsung BAE
PY - 2005
DO - 10.1093/ietisy/e88-d.3.435
JO - IEICE TRANSACTIONS on Information
SN -
VL - E88-D
IS - 3
JA - IEICE TRANSACTIONS on Information
Y1 - March 2005
AB - In this paper, we propose a new feature extraction method that combines both HMT-based denoising and weighted filter bank analysis for robust speech recognition. The proposed method is made up of two stages in cascade. The first stage is denoising process based on the wavelet domain Hidden Markov Tree model, and the second one is the filter bank analysis with weighting coefficients obtained from the residual noise in the first stage. To evaluate performance of the proposed method, recognition experiments were carried out for additive white Gaussian and pink noise with signal-to-noise ratio from 25 dB to 0 dB. Experiment results demonstrate the superiority of the proposed method to the conventional ones.
ER -