A Novel Teaching Video Speech Recognition Method Based on HMM Model

Abstract

To facilitate users in information retrieval based on teaching video content, a teaching video recognition training system with waveform display, pronunciation evaluation and other functions is developed by using speech recognition technology. In this paper, the basic principle of speech recognition is described in detail from the aspects of speech signal preprocessing, feature parameter extraction and HMM model matching. HMM, Viterbi decoding and speech evaluation algorithm are designed and applied. Then, based on the online collaborative learning platform, the recognition model is established by HMM, and the training resources in the experimental database are tested and analyzed. The results show that the improved speech recognition model can enhance the robustness and efficiency of the system, and improve the accuracy and error correction rate of English video teaching score.