基于语音识别的汉语发音自动评分系统的设计与实现

Design and implementation of automatic Chinese speech scoring system based on speech recognition

语音识别技术的发展使得人与计算机的交互成为可能,针对目前对外汉语中发音教学的不足,在结合了语音识别的相关原理,提出了在对外汉语教学领域中汉语自动发音水平评价系统的设计,详细地描述了系统的结构、功能及流程。介绍了系统实现中的关键技术和步骤:动态时间弯折算法、语料库的建立、声韵分割技术以及评价分级标准。通过小范围的试验,表明该系统对留学生汉语发音水平的测试有一定的参考价值。

The technology of speech recognition enabled the interaction between human and computer. A design of Chinese speech scoring system in the field of teaching Chinese as a foreign language combined with some technology of speech recognition is put forward. The architecture, function, flow of the system are expounded. Some key technology and steps such as dynamic time warping, corpora establishment, Initial/Final segmentation and gradation criteria are introduced. The result of the test shows the system is helpful for evaluation of foreign students＇ Chinese pronunciation.