基于语音识别技术的英语口语自学系统评分机制的研究

The Research of the Scoring Mechanism of Spoken English Learning System Based on the Technology of Speech Recognition

在外语学习中，计算机辅助发音学习系统给学习者提供有效的发音指导。该文结合语音识别的相关原理，研究英语口语自学系统评分机制的关键技术，针对目前发展状况提出语音评分机制存在的问题，并从主观和客观两方面入手对多种语音评分技术的优缺点进行比较，提出将评分分为三大部分，即声学评分、韵律评分和感知评分，将主观评价和客观评价整合起来，最终提出基于HMM技术和神经网络技术的评分机制，促进英语口语自学能力的提高。

In the process of the foreign language study, computer-assisted language learning system provides effective guidance for pronunciation. Combined with the principles of the speech recognition, the key issue of the scoring mechanism for the spoken English self-learning system is investigated in this paper. Based on the current status, the advantage and the disadvantage of the scoring techniques are discussed from the subjective and objective aspects. To integrate subjective and objective evaluation, we separate the scoring into three parts, such as acoustics scoring, rhythm scoring and perception scoring, the scoring mechanism based upon the HMM technique and neural network technique is proposed to promote the self-learning ability of English.