Design of Oral English Teaching Assistant System for Students Based on Virtual Reality and Artificial Intelligence Algorithm

Cui Li

DOI: https://doi.org/10.1109/ICTEI60496.2023.00067

2023-09-11

Abstract:With the increasing importance of spoken English, it can no longer meet the needs of the times to score spoken English pronunciation by hand. VR (virtual reality) has become a new development direction, which can realize remote control and resource sharing, and has a good development prospect. In order to improve the scoring efficiency, this paper proposes a student oral English teaching assistant system based on VR and AI (Artificial Intelligence) algorithm. This system can be divided into six modules, namely, teacher module, student module, learning statistics module, administrator module, display module and oral evaluation module, and each functional module is complementary to each other. The system uses a speaker-independent oral pronunciation evaluation method combining spectrogram and CNN (Convolutional Neural Network). The speech samples are used as a standard sample training set, and then are sent to the network model for training, so as to obtain the standard pronunciation spectrogram of each phoneme level; Then, any input speech is taken as the test result of the test set. If the output phoneme is inconsistent with the expectation, it is considered as nonstandard, and the phoneme-specific evaluation can be given. The research results show that the scoring accuracy of this system is better than that of the traditional system, and the scoring efficiency is higher, which shows that the performance of this system is better. This system solves the problems of low timeliness, complicated operation process and low scoring accuracy of the traditional system, and provides a guarantee for the oral English pronunciation quality scoring system.

Computer Science,Education