English Speech Recognition Method Based on HMM Technology

In order to solve the problem of insufficient feedback and intelligent speech recognition in English pronunciation learning, this paper designs and implements an English speech learning system based on speech recognition technology. Matlab is used to complete the sample speech data collection and preprocessing. Combined with the needs of Chinese speaking English learners, the speech recognition technology based on HMM model is used to decode the speech by Viterbi, and the recognition score is obtained by posterior probability. The experimental results show that this scheme greatly reduces the search space and effectively shortens the algorithm time without affecting the recognition accuracy. The function of each module of the system has been well implemented. The system can be used by users normally and is stable in function and performance.