PASCAL - Pattern Analysis, Statistical Modelling and Computational Learning

Combined Waveform-Cepstral Representationfor Robust Speech Recognition
Matthew Ager, Zoran Cvetkovic and Peter Sollich
In: 2011 IEEE International Symposium on Information Theory, 31 July - 5 Aug 2011, Russia.

Abstract

High-dimensional acoustic waveform representations are studied as a front-end for noise robust automatic speech recognition using generative methods, in particular Gaussian mixture models and hidden Markov models. The proposed representations are compared with standard cepstral features on phoneme classification and recognition tasks. While lower error rates are achieved using cepstral features at very low noise levels, the models using acoustic waveform representations are more robust to additive noise. A combination of acoustic waveforms and cepstral features achieves better results than either of the individual representations across all noise levels.

PDF - PASCAL Members only - Requires Adobe Acrobat Reader or other PDF viewer.
EPrint Type:Conference or Workshop Item (Paper)
Project Keyword:Project Keyword UNSPECIFIED
Subjects:Speech
ID Code:8001
Deposited By:Matthew Ager
Deposited On:17 March 2011