Can back-ends be more robust than front-ends? Investigation over the Aurora-2 database

Author

Bernard, Alexis ; Gong, Yifan ; Cui, Xiaodong

Author_Institution

Speech Technol. Lab., Texas Instrum. Inc, Dallas, TX, USA

Volume

1

fYear

2004

fDate

17-21 May 2004

Abstract

We present a back-end solution developed at Texas Instruments for noise robust speech recognition. The solution consists of three techniques: 1) a joint additive and convolutive noise compensation (JAC) which adapts speech acoustic models; 2) an enhanced channel estimation procedure which extends JAC performance towards lower SNR ranges; 3) an N-pass decoding algorithm. The performance of the proposed back-end is evaluated on the Aurora-2 database. With 20% fewer model parameters and without the need for the second order derivative of the recognition features, the performance of the proposed solution is 91.86%, which outperforms that of the ETSI advanced front-end standard (88.19%) by more than 30% relative word error rate reduction.

Keywords

acoustic noise; adaptive signal processing; channel estimation; decoding; error statistics; random noise; speech recognition; Aurora-2 database; ETSI advanced front-end standard; additive distortion compensation; additive noise compensation; back-end solution; channel estimation; convolutive distortion compensation; convolutive noise compensation; decoding algorithm; noise robust speech recognition; second order derivative; speech acoustic model adaptation; word error rate reduction; Acoustic noise; Additive noise; Channel estimation; Databases; Decoding; Instruments; Noise robustness; Signal to noise ratio; Speech enhancement; Speech recognition;

fLanguage

English

Publisher

ieee

Conference_Titel

Acoustics, Speech, and Signal Processing, 2004. Proceedings. (ICASSP '04). IEEE International Conference on

ISSN

1520-6149

Print_ISBN

0-7803-8484-9

Type

conf

DOI

10.1109/ICASSP.2004.1326163

Filename

1326163