Method and apparatus for a robust feature extraction for speech recognition

US6678657B1 · kind B1 · utility

Assignee

• Telefonaktiebolaget LM Ericsson (Publ) · SE

Inventors

• Raymond Brückner · Blaustein, DE
• Hans-Gunter Hirsch · Heroldsberg, DE
• Rainer Klisch · Koblenzer Straße, DE
• Volker Springer · Wendelstein, DE

Key dates

• Filing date: Oct 23, 2000
• Grant date: Jan 13, 2004
• Expiry date: Feb 15, 2022

Classification

• Technology area (CPC G): Physics
• CPC primary: G10L21/0208
• WIPO field: Computer technology
• WIPO sector: Electrical engineering

Abstract

The present invention relates to a method and an apparatus for a robust feature extraction for speech recognition in a noisy environment, wherein the speech signal is segmented and is characterized by spectral components. The speech signal is splitted into a number of short term spectral components in L subbands, with L=1, 2, . . . and a noise spectrum from segments that only contain noise is estimated. Then a spectral subtraction of the estimated noise spectrum from the corresponding short term spectrum is performed and a probability for each short term spectrum component to contain noise is calculated. Finally these spectral component of each short-term spectrum, having a low probability to contain speech are interpolated in order to smooth those short-term, spectra that only contain noise. With the interpolation the spectral components containing noise are interpolated by reliable spectral speech components that could be found in the neighborhood.