A cepstral noise reduction multi-layer neural network

Author

Sorensen, Helge B D

Author_Institution

Inst. of Electron. Syst., Aalborg Univ., Denmark

fYear

1991

fDate

14-17 Apr 1991

Firstpage

933

Abstract

The problem of speech recognition in the presence of interfering nonstationary noise is addressed. A method for noise reduction in the cepstral domain based on a multilayer network is proposed and tested on a large database of isolated words contaminated with nonstationary F-16 jet noise. The speech recognition system consists of an auditory preprocessing module, the cepstral noise reduction multilayer network, and a neural network classifier. The noise reduction network performs a nonlinear autoassociative mapping in the cepstral domain between a set of noisy cepstral coefficients and a set of noise-free cepstral coefficients. The average recognition rate on a test database was improved up to 65% when the noise reduction network was added to the speech recognition system

Keywords

interference suppression; neural nets; noise; speech recognition; F-16 jet noise; auditory preprocessing module; cepstral noise reduction; interfering nonstationary noise; isolated words; large database; multi-layer neural network; neural network classifier; noise-free cepstral coefficients; noisy cepstral coefficients; nonlinear autoassociative mapping; recognition rate; speech recognition; Cepstral analysis; Databases; Multi-layer neural network; Neural networks; Noise reduction; Signal mapping; Speech enhancement; Speech recognition; Testing; White noise;

fLanguage

English

Publisher

ieee

Conference_Titel

Acoustics, Speech, and Signal Processing, 1991. ICASSP-91., 1991 International Conference on

Conference_Location

Toronto, Ont.

ISSN

1520-6149

Print_ISBN

0-7803-0003-3

Type

conf

DOI

10.1109/ICASSP.1991.150493

Filename

150493