Investigations of HMM-Based Speech Recognition Technology

Abstract

We are in an era of informatization and big data. In this context, speech data are increasing and becoming more and more frequently used. When we choose and use a lot of speech information on the Internet, speech recognition technology shows its great importance. In this paper, by studying speech signal preprocessing, HMM model, MFCC feature parameter extraction and other technologies, a speech recognition algorithm based on HMM is derived to achieve a more accurate result of speech words and phrases.