A Time Reduction Algorithm Using Vowel Classification for Large-Vocabulary Speech Recognition

Abstract

In this paper, a time reduction algorithm for large-vocabulary speech recognition has been studied. To reduce the recognition time, vowel was classified according to its formant information based on linear prediction analysis. Formant extraction was done in the assumed vowel regions. To segment vowel regions we used acoustic features such as energy, zero crossing rate and filter bank outputs. Computer simulation was done to test the time reduction algorithm with 1160 words. Simulation results show that the recognition rate is about 97% and the time for recognition can be reduced by 80% as compared to the case without the reduction algorithm.