ASR in mobile phones - an industrial approach

Author

Varga, Imre ; Aalburg, Stefanie ; Andrassy, Bernt ; Astrov, Sergey ; Bauer, Josef G. ; Beaugeant, Christophe ; Geissler, C. ; Höge, Harald

Author_Institution

Siemens AG, Munich, Germany

Volume

10

Issue

8

fYear

2002

fDate

11/1/2002 12:00:00 AM

Firstpage

562

Lastpage

569

Abstract

In order to make hidden Markov model (HMM) speech recognition suitable for mobile phone applications, Siemens developed a recognizer, Very Smart Recognizer (VSR), for deployment in future mobile phone generations. Typical applications will be name dialling, command and control operations suited for different environments, for example in cars. The paper describes research and development issues of a speech recognizer in mobile devices focusing on noise robustness, memory efficiency and integer implementation. The VSR is shown to reach a word error rate as low as 4.1% on continuous digits recorded in a car environment. Furthermore by means of discriminative training and HMM-parameter coding, the memory requirements of the VSR HMMs are smaller than 64 kBytes.

Keywords

hidden Markov models; mobile handsets; speech recognition; DSP implementation; HMM-parameter coding; Siemens; Very Smart Recognizer; automatic speech recognition; cars; command and control; discriminative training; hidden Markov model; integer implementation; memory efficiency; memory requirements; mobile devices; mobile phones; name dialling; noise robustness; research and development; word error rate; Automatic speech recognition; Command and control systems; Error analysis; Hidden Markov models; Mobile handsets; Noise robustness; Research and development; Speech enhancement; Speech recognition; Working environment noise;

fLanguage

English

Journal_Title

Speech and Audio Processing, IEEE Transactions on

Publisher

ieee

ISSN

1063-6676

Type

jour

DOI

10.1109/TSA.2002.804548

Filename

1175528