Title :
Noise compensation methods for hidden Markov model speech recognition in adverse environments
Author :
Vaseghi, Saeed V. ; Milner, Ben P.
Author_Institution :
Dept. of Electr. & Electron. Eng., Queen´´s Univ., Belfast, UK
fDate :
1/1/1997 12:00:00 AM
Abstract :
Several noise compensation schemes for speech recognition in impulsive and nonimpulsive noise are considered. The noise compensation schemes are spectral subtraction, HMM-based Wiener (1949) filters, noise-adaptive HMMs, and a front-end impulsive noise removal. The use of the cepstral-time matrix as an improved speech feature set is explored, and the noise compensation methods are extended for use with cepstral-time features. Experimental evaluations, on a spoken digit database, in the presence of ear noise, helicopter noise, and impulsive noise, demonstrate that the noise compensation methods achieve substantial improvement in recognition across a wide range of signal-to-noise ratios. The results also show that the cepstral-time matrix is more robust than a vector of identical size, which is composed of a combination of cepstral and differential cepstral features
Keywords :
Wiener filters; acoustic noise; adaptive signal processing; cepstral analysis; feature extraction; filtering theory; hidden Markov models; noise abatement; speech processing; speech recognition; HMM based Wiener filters; adverse environments; cepstral features; cepstral-time matrix; differential cepstral features; ear noise; experimental evaluations; front-end impulsive noise removal; helicopter noise; hidden Markov model; impulsive noise; noise adaptive HMM; noise compensation methods; nonimpulsive noise; signal to noise ratios; spectral subtraction; speech feature set; speech recognition; spoken digit database; Cepstral analysis; Ear; Helicopters; Hidden Markov models; Noise robustness; Signal to noise ratio; Spatial databases; Speech enhancement; Speech recognition; Wiener filter;
Journal_Title :
Speech and Audio Processing, IEEE Transactions on
