抽象的な

The Learning Method of Speech Recognition Based on HMM

Vidwath R Hebse , Anitha G

This paper helps to improve the student’s interest towards speech recognition. It implements Automatic Speech Recognition System based on the following: Preprocessing, Feature Extraction Technique (MFCC- Mel Frequency Cepstrum Coefficients) and Hidden Markov Model (used in recognition phase) .Its purpose is that the human voice can be converted to a computer-readable input, Such as buttons, binary-coded, or sequence of characters. Acoustic phonetic approach, Knowledge based approach and Pattern recognition approach are the three approaches used in speech recognition. In this paper Pattern recognition technique is used. This approach requires no explicit knowledge of speech. This approach has two steps – namely, training of speech patterns based on some generic spectral parameter set and recognition of patterns via pattern comparison.