Optimizing feature extraction for speech recognition

被引：20

作者：

Lee, CH ^{[1
]}

Hyun, DH ^{[1
]}

Choi, ES ^{[1
]}

Go, JW ^{[1
]}

Lee, CY ^{[1
]}

机构：

[1] Yonsei Univ, Dept Elect & Elect Engn, Seoul 120749, South Korea

来源：

IEEE TRANSACTIONS ON SPEECH AND AUDIO PROCESSING | 2003年 / 11卷 / 01期

关键词：

critical band filters; feature extraction; melcepstrum; optimization; speech recognition;

D O I：

10.1109/TSA.2002.805644

中图分类号：

O42 [声学];

学科分类号：

070206 ; 082403 ;

摘要：

In this paper, we propose a method to minimize the loss of information during the feature extraction stage in speech recognition by optimizing the parameters of the mel-cepstrum transformation, a transform which is widely used in speech recognition. Typically, the mel-cepstrum is obtained by critical band filters whose characteristics play an important role in converting a speech signal into a sequence of vectors. First, we analyze the performance of the mel-cepstrum by changing the parameters of the filters such as shape, center frequency, and bandwidth. Then we propose an algorithm to optimize the parameters of the filters using the simplex method. Experiments with Korean digit words show that the recognition rate improved by about 4-7%.

引用

页码：80 / 87

页数：8

共 17 条

[1]

Biem A, 1997, INT CONF ACOUST SPEE, P1503, DOI 10.1109/ICASSP.1997.596235

[2]

BIEM A, 1994, INT CONF ACOUST SPEE, P485

[3]

Buchanan J., 1992, NUMERICAL METHODS AN

[4] Speech feature extracted from adaptive wavelet for speech recognition [J].

Chang, SW ;

Kwon, Y ;

Yang, SI .

ELECTRONICS LETTERS, 1998, 34 (23) :2211-2213

[5] HMM-based speech recognition using state-dependent, discriminatively derived transforms on mel-warped DFT features [J].

Chengalvarayan, R ;

Deng, L .

IEEE TRANSACTIONS ON SPEECH AND AUDIO PROCESSING, 1997, 5 (03) :243-256

[6]

Deller Jr J. R., 1993, DISCRETE TIME PROCES

[7]

Fukunaga K., 1990, INTRO STAT PATTERN R

[8]

Gopinath RA, 1998, INT CONF ACOUST SPEE, P661, DOI 10.1109/ICASSP.1998.675351

[9]

GU L, 1996, P INT C SIGN PROC, P745

[10] Analysis of LPC/DFT features for an HMM-based alphadigit recognizer [J].

Mashao, DJ ;

Gotoh, Y ;

Silverman, HF .

IEEE SIGNAL PROCESSING LETTERS, 1996, 3 (04) :103-106

← 1 2 →