Design of English Oral Speech Correction System Based on Deep Neural Networks
Caiyun Li, Jianxiang Wang, Yan Wang · 2024
As one of the three major language elements, phonetics plays a crucial role in English learning and is also a key link in English teaching. The effectiveness of its teaching can directly affect the success of English teaching. This English oral speech correction system uses machine speech scoring method to score the speech of the subjects. By comparing the speech of the subjects with standard speech, the system uses frame based windowing and endpoint detection methods to preprocess the speech of the subjects. The Mel frequency cepstral coefficient method was used to extract corresponding feature intervals from speech, and the dynamic time warping method was used to dynamically induce and classify them. By correcting the resonance peak image, the average effective rate of vowels, consonants, and words was 93.91%, and the system could effectively correct speech.