UEP-driven extended feature extraction for ASR over 3G speech channels

Author

Gomez-Cajas, D.F. ; Pelaez-Moreno, C. ; Diaz-de-Maria, Fernando

Author_Institution

Dept. of Signal Theor. & Commun., Univ. Carlos III of Madrid, Leganes, Spain

fYear

2012

fDate

1-2 Nov. 2012

Firstpage

1

Lastpage

5

Abstract

In this paper we present an extended feature extraction procedure for Automatic Speech Recognition (ASR) over 3G UMTS channels [2], within a bitstream-based Network Speech Recognition (NSR) architecture. This procedure takes advantage of the Unequal Error Protection (UEP) policy that is applied by the channel coder that highly protects selected parts of the transmitted information, leaving some parameters almost unprotected. Given that wireless cellular channels are very prone to errors, taking this fact into account, the new front-end produces important improvements in the performance of recognizers, increasing the benefits as the quality of channel worsens.

Keywords

3G mobile communication; channel coding; feature extraction; speech coding; speech recognition; 3G UMTS channels; 3G speech channels; ASR; UEP-driven extended feature extraction; automatic speech recognition; bitstream-based NSR architecture; bitstream-based network speech recognition architecture; channel coder; unequal error protection policy; wireless cellular channels; 3G mobile communication; Feature extraction; Robustness; Speech; Speech coding; Speech recognition; Vectors; UMTS; automatic speech recognition; channel coding; feature extraction; network speech recognition;

fLanguage

English

Publisher

ieee

Conference_Titel

Circuits and Systems (CWCAS), 2012 IEEE 4th Colombian Workshop on

Conference_Location

Barranquilla

Print_ISBN

978-1-4673-4612-2

Type

conf

DOI

10.1109/CWCAS.2012.6404082

Filename

6404082