English Speech Recognition Method Based on Hidden Markov Model

This paper aims to tackle the problem of English speech recognition, which is to seek the most suitable word sequence given a segment of English voice. In this paper, the English speech recognition system is made up of four parts, that is, 1) Voice acquisition, 2) Speech model, 3) Speech recognition and 4) Speech recognition results. Main idea of this paper is to identify English speech with Hidden Markov model. To enhance performance of the HMM, we discuss how to optimize parameters in HMM. To demonstrate the effectiveness of our method, The Aurora 2 English language database is utilized. Then, we test the English speech recognition accuracy of four different noisy environments. Experimental results prove that the proposed approach can effectively enhance accuracy of English speech recognition process.