Research on Speech Recognition Technology of Oral English Learning Based on Improved GLR Algorithm

Abstract

At present, the language teaching software lacks good feedback on the results of oral English learning, and the most important thing in oral English learning is to evaluate it effectively. In this paper, according to the pronunciation characteristics of English and the needs of spoken English learners, the speech recognition process of spoken English is realized by using the improved GLR (Generalized Maximum Likelihood Ratio Detection) algorithm, and a pronunciation scoring system for spoken English practice is provided for learners to help them improve their pronunciation in time. Comprehensive evaluation results show that the improved GLR algorithm improves the operation speed and processing performance compared with statistical algorithm and dynamic memory algorithm, which is more suitable for the task of speech recognition in oral English learning, and provides a new idea in the field of speech recognition in oral English learning.