Specifics of hidden Markov model modifications for large vocabulary continuous speech recognition

被引：1

作者：

Silingas, D

Telksnys, L

机构：

[1] Vytautas Magnus Univ, Dept Appl Informat, LT-3035 Kaunas, Lithuania

[2] Vytautas Magnus Univ, Recognit Proc Dept, Inst Math & Informat, Dept Appl Informat, LT-08663 Vilnius, Lithuania

来源：

INFORMATICA | 2004年 / 15卷 / 01期

关键词：

large vocabulary continuous speech recognition; hidden Markov model; Viterbi recognition; beam search; context-dependent phones; Gaussian mixture; language modeling; HTK; WSJCAM0;

D O I：

暂无

中图分类号：

TP [自动化技术、计算机技术];

学科分类号：

0812 ;

摘要：

Specifics of hidden Markov model-based speech recognition are investigated. Influence of modeling simple and context-dependent phones, using simple Gaussian, two and three-component Gaussian mixture probability density functions for modeling feature distribution, and incorporating language model are discussed. Word recognition rates and model complexity criteria are used for evaluating suitability of these modifications for practical applications. Development of large vocabulary continuous speech recognition system using HTK toolkit and WSJCAMO English speech corpus is described. Results of experimental investigations are presented.

引用

页码：93 / 110

页数：18

共 18 条

[1]

[Anonymous], 2002, CAMBRIDGE U ENG DEP

[2]

[Anonymous], SURVEY STATE ART HUM

[3]

[Anonymous], CUEDFINFENGTR38

[4]

CLARKSON PR, 1997, P EUROSPEECH 97

[5]

Deller J.R., 1993, Discrete-time processing of speech signals

[6] A statistical text-to-phone function using ngrams and rules [J].

Fisher, WM .

ICASSP '99: 1999 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH, AND SIGNAL PROCESSING, PROCEEDINGS VOLS I-VI, 1999, :649-652

[7]

FRANSEN J, 1994, CUEDINFENGTR192

[8]

GAUVAIN JL, 1996, T IEICE

[9]

GAUVAIN JL, 1994, LIMSI NOV 93 WSJ SYT

[10]

Jurafsky D., 2000, Speech and Language Processing. An Introduction to Natural language Processing, Computational Linguistics

← 1 2 →