Neural dual extended Kalman filtering: applications in speech enhancement and monaural blind signal separation

Author

Wan, Eric A. ; Nelson, Alex T.

Author_Institution

Dept. of Electr. Eng., Oregon Graduate Inst., Portland, OR, USA

fYear

1997

fDate

24-26 Sep 1997

Firstpage

466

Lastpage

475

Abstract

The removal of noise from speech signals has applications ranging from speech enhancement for cellular communications, to front ends for speech recognition systems. A nonlinear time-domain method called dual extended Kalman filtering (DEKF) is presented for removing nonstationary and colored noise from speech. We further generalize the algorithm to perform the blind separation of two speech signals from a single recording

Keywords

Kalman filters; autoregressive processes; filtering theory; neural nets; noise; nonlinear filters; speech enhancement; state estimation; state-space methods; cellular communications; colored noise; monaural blind signal separation; neural dual extended Kalman filtering; nonlinear time-domain method; nonstationary noise; speech enhancement; speech recognition systems; speech signals; Additive noise; Filtering; Integrated circuit modeling; Kalman filters; Neural networks; Signal processing; Speech enhancement; Speech processing; State estimation; Time domain analysis;

fLanguage

English

Publisher

ieee

Conference_Titel

Neural Networks for Signal Processing [1997] VII. Proceedings of the 1997 IEEE Workshop

Conference_Location

Amelia Island, FL

ISSN

1089-3555

Print_ISBN

0-7803-4256-9

Type

conf

DOI

10.1109/NNSP.1997.622428

Filename

622428