Duration modeling in large vocabulary speech recognition

Author

Anastasakos, Anastasios ; Schwartz, Richard ; Shu, Han

Author_Institution

Northeastern Univ., Boston, MA, USA

Volume

1

fYear

1995

fDate

9-12 May 1995

Firstpage

628

Abstract

This paper presents a study of different methods for phoneme duration modeling in large vocabulary speech recognition. We investigate the employment of phoneme duration and the effect of context, speaking rate and lexical stress in the duration of phoneme segments in a large vocabulary speech recognition system. The duration models are used in a postprocessing phase of BYBLOS, our baseline HMM-based recognition system, to rescore the N-Best hypotheses. We describe experiments with the 5 K word ARPA Wall Street Journal (WSJ) corpus. The results show that integration of duration models that take into account context and speaking rate can improve the word accuracy of the baseline recognition system

Keywords

hidden Markov models; speech recognition; ARPA Wall Street Journal corpus; BYBLOS; N-Best hypotheses; baseline HMM-based recognition system; context; duration modeling; large vocabulary speech recognition; lexical stress; phoneme; postprocessing phase; speaking rate; word accuracy; Context modeling; Costs; Density functional theory; Employment; Hidden Markov models; Signal analysis; Speech analysis; Speech recognition; Stress; Vocabulary;

fLanguage

English

Publisher

ieee

Conference_Titel

Acoustics, Speech, and Signal Processing, 1995. ICASSP-95., 1995 International Conference on

Conference_Location

Detroit, MI

ISSN

1520-6149

Print_ISBN

0-7803-2431-5

Type

conf

DOI

10.1109/ICASSP.1995.479676

Filename

479676