Home

About

Advanced Search

Browse by Discipline

Scientific Societies

E-print Alerts

Add E-prints

E-print Network
FAQHELPSITE MAPCONTACT US


  Advanced Search  

 
Non-linear feature extraction for robust speech recognition in stationary and non-stationary noise q
 

Summary: Non-linear feature extraction for robust speech recognition
in stationary and non-stationary noise q
Qifeng Zhu, Abeer Alwan *
Department of Electrical Engineering, The Henry Samuli School of Engineering and Applied Science, 66-147E Engr. IV,
UCLA 405 Hilgard Avenue, Box 951594, Los Angeles, CA 90095-1594, USA
Received 9 August 2001; received in revised form 22 October 2002; accepted 22 March 2003
Abstract
An analysis-based non-linear feature extraction approach is proposed, inspired by a model of how
speech amplitude spectra are affected by additive noise. Acoustic features are extracted based on the noise-
robust parts of speech spectra without losing discriminative information. Two non-linear processing
methods, harmonic demodulation and spectral peak-to-valley ratio locking, are designed to minimize
mismatch between clean and noisy speech features. A previously studied method, peak isolation [IEEE
Transactions on Speech and Audio Processing 5 (1997) 451], is also discussed with this model. These
methods do not require noise estimation and are effective in dealing with both stationary and non-sta-
tionary noise. In the presence of additive noise, ASR experiments show that using these techniques in the
computation of MFCCs improves recognition performance greatly. For the TI46 isolated digits database,
the average recognition rate across several SNRs is improved from 60% (using unmodified MFCCs) to 95%
(using the proposed techniques) with additive speech-shaped noise. For the Aurora 2 connected digit-string
database, the average recognition rate across different noise types, including non-stationary noise back-
ground, and SNRs improves from 58% to 80%.

  

Source: Alwan, Abeer - Electrical Engineering Department, University of California at Los Angeles

 

Collections: Computer Technologies and Information Sciences