• TITLE: SIGNAL PROCESSING TECHNIQUES FOR SPEECH RECOGNITION
• AUTHOR(S): R. T. Ilarionov, T. D. Todorov, G. S. Tsanev, S. Y. Yordanov
• ABSTRACT: Most of the authors analyzing speech use the wavelet method for feature extraction and Dynamic Time Warping for comparing two time series in their classifiers. The usage of DTW generates significant difficulties obtaining pattern as an average of series. We introduce an alternative approach without analysis of the cepstral coefficients. The present paper summarizes new methods for recognition of speech. A new finite element signal processing method is demonstrated. The present speech recognition techniques essentially reduce the total computational work, which is very important in the case of a real time application. Successful real life examples are presented. The tests strongly support the considered theory.
• DOWNLOAD: SIGNAL PROCESSING TECHNIQUES FOR SPEECH RECOGNITION