The Use of Wide-frame Cepstrum in Speech Recognition
The paper presents a new approach to speech recognition using cepstral analysis. Cepstrum is being counted in just two wide frames for each syllable in word. The algorithm is discussed and example of application using backpropagation neural network is given.
KeywordsSpeech Recognition Excitation Signal Phoneme Recognition Wide Frame Polish Word
Unable to display preview. Download preview PDF.
- Brze§cinska J, de Jastrz@biec Wykowski T. “Computer Usage’s Support for Disabled”, Metallurgy and Foundry Engineering — Special Edition, Krakow 2001, pp. 99–106 (in Polish)Google Scholar
- Deller J. Jr., Hansen J., Proakis J.“Discrete — Time Processing of Speech Signal”, IEEE Press, New York 2000Google Scholar
- Grocholewski S. “Statistical Foundations of ARM System For Polish Language”, Poznan University of Technology Printing House, Poznan 2001 (in Polish)Google Scholar
- Oppenheim A.V. and al. “Applications of Digital Signal Processing”, Prentice-Hall, Inc. Englewood Cliffs, New Jersey USA 1978Google Scholar
- .Tadeusiewicz R. „Speech Signals“, Transport and Communication Publishers, Warsaw 1988 (in Polish)Google Scholar