Speech Recognition Model Based on Deep Learning And Application in Pronunciation Quality Evaluation System
Teng Haikun, Shiying Wang, Liu Xinsheng, Gabriel Xiao-Guang Yue · 2019
In order to improve the performance of the speech recognition and pronunciation quality evaluation system, the deep learning-based computer aided foreign language pronunciation evaluation and learning system has become a research focus of current artificial intelligence technology. Combining with the current advanced voice information technology theory, based on the previous research results, proposed to sparse since the encoder deep learning neural network is applied to speech recognition, using sparse automatic encoder based on MFCC features in-depth study, the imitation of the auditory nerve sparse touches the depth of feature extracting signal, beneficial to the improvement of the HMM model for speech recognition accuracy, meet the needs of the current computer assisted English teaching. The simulation results show that the recognition rate of the deep learning neural network is obviously superior to that of the traditional speech recognition algorithm, which realizes more accurate human-computer interaction and improves the reliability of the evaluation of the quality of foreign language pronunciation.