APPARATUS FOR RECOGNIZING VOICE USING HMM MODEL SIMILARITY AND RECOGNIZING METHOD THEREOF

摘要

PURPOSE: An apparatus for recognizing voice using HMM(hidden markov model) model similarity and a recognizing method thereof are provided to calculate the similarity between vocabularies without using viterbi algorithm. CONSTITUTION: An apparatus for recognizing voice using HMM(hidden markov model) model similarity includes a microphone(11), a low pass filter(13), an analog/digital converter(15), a microcomputer(17), a voice recognizer(19) and a driver(21). The microphone(11) converts voice signal into electrical signal. The low pass filter(13) receives the signal from the microphone and eliminates the high frequency component. The analog/digital converter(15) converts the analog signal to digital signal for the microcomputer. The microcomputer(17) generates control signal for the driver. The voice recognizer(19) discriminates voices at vocabulary model basis. The driver(21) outputs control signal to the microcomputer.