English Real-time Speech Recognition Based on Hidden Markov and Edge Computing Model
Juan Wu · 2021 Third International Conference on Inventive Research in Computing Applications (ICIRCA) · 2021
This paper is based on the hidden Markov model and edge computing to recognize English speech in real time. This article first analyzes the preprocessing process of the voice signal, and proposes the use of dual-threshold method to detect the voice signal endpoint. Then, two characteristic parameter extraction methods are analyzed in detail, including linear prediction cepstral coefficient and Mel cepstral coefficient. Finally, this paper focuses on the real-time recognition algorithm of English speech based on HMM and edge computing. Experiments have proved the effectiveness and feasibility of this algorithm.