Noise robust speech parameterization using multiresolution feature extraction.

In this paper, we present a multiresolution-based feature extraction technique for speech recognition in adverse conditions. The proposed front-end algorithm uses mel cepstrum-based feature computation of subbands in order not to spread noise distortions over the entire feature space. Conventional f...

Full description

Bibliographic Details
Published in:	IEEE Transactions on speech and audio processing 9, 8 (2001).
Main Author:	Hariharan, R.
Format:	Article
Language:	English
Subjects:	HMM parameters. PCA transform. SNR. Cepstral domain feature vector normalization. Computational complexity reduction. Connected digit recognition task. Feature vector dimension. Front-end algorithm. Long-term spectral information. Mel cepstrum-based feature computation. Multiresolution-based feature extraction. Noise conditions. Noise robustness. Speech parameterization. Speech recognition. Subbands full-band features. Word recognition accuracy.