Solutions for robust recognition over the GSM cellular network

Author

Karray, Lamia ; Jelloun, Abdellatif Ben ; Mokbel, Chafic

Author_Institution

CNET, Lannion, France

Volume

1

fYear

1998

fDate

12-15 May 1998

Firstpage

261

Abstract

This paper deals with automatic speech recognition robustness for noisy wireless communications. We propose several solutions to improve speech recognition over the cellular network. Two architectures are derived for the recognizer. They are based on hidden Markov models (HMMs) adapted to adverse noise conditions. Then two more specific solutions aiming to alleviate GSM cellular network defects (holes and impulsive noise) are developed. Holes are detected and rejected. Impulsive noises are modeled using mixture density HMMs and a maximum likelihood criterion. These solutions allow a noticeable recognition error reduction. The last one seems to be promising

Keywords

Gaussian distribution; cellular radio; hidden Markov models; land mobile radio; maximum likelihood estimation; noise; radio networks; speech recognition; GSM cellular network; Gaussian noise modelling; automatic speech recognition; hidden Markov models; holes detection; holes rejection; impulsive noise; maximum likelihood criterion; mixture density HMM; multi-Gaussian distribution; noisy wireless communications; recognition error reduction; robust recognition; Acoustic noise; Databases; GSM; Hidden Markov models; Land mobile radio cellular systems; Noise robustness; Speech enhancement; Speech recognition; Vocabulary; Working environment noise;

fLanguage

English

Publisher

ieee

Conference_Titel

Acoustics, Speech and Signal Processing, 1998. Proceedings of the 1998 IEEE International Conference on

Conference_Location

Seattle, WA

ISSN

1520-6149

Print_ISBN

0-7803-4428-6

Type

conf

DOI

10.1109/ICASSP.1998.674417

Filename

674417