摘要
汉英口语翻译自动评分在确保评分客观性、提高评分效率、降低测试成本方面发挥着至关重要作用,因此,基于语音信号处理和文本识别技术研究了汉英口语翻译自动评分方法。先对语音信号进行文本识别,计算相似度特征;接着对语音信号降噪预处理,计算降噪后信号语音特征,最终得到包含6个特征的自动评分系统。以某大学外国语学院汉英口语翻译所采集的语音信号为例进行分析,得到了包含4个特征的自动评分系统。结果表明,在去除冗余特征后,自动评分系统的性能得到了明显改善,这对汉英口语翻译自动评分具有参考价值。
Automatic scoring of Chinese-English oral translation plays an important role in ensuring the objectivity of scoring,improving scoring efficiency and reducing test cost.This paper studies the automatic scoring method of Chinese-English oral translation based on speech signal processing and text recognition technology.The text recognition of speech signal is carried out,and the similarity feature is calculated;the speech signal denoising is preprocessed,and the speech feature after denoising is calculated,and finally the automatic scoring system including six features is obtained.Based on the analysis of the speech signals collected by the college of foreign languages of a university,an automatic scoring system with four features is obtained.The results show that the performance of the automatic scoring system is significantly improved after removing the redundant features,which has a certain reference value for the automatic scoring of Chinese-English oral translation.
作者
汪斐
王婧锦
WANG Fei;WANG Jingjin(School of Humanities, Shangluo University, Shangluo 726000, China;School of Foreign Languages, Shaanxi University of Traditional Chinese Medicine, Xianyang 712046, China)
出处
《微型电脑应用》
2021年第10期39-41,共3页
Microcomputer Applications
基金
陕西省教育科学“十三五”规划课题(SGH18H400)。
关键词
文本识别
语音信号降噪
自动评分
相似度特征
语音特征
text recognition
noise reduction of speech signal
automatic scoring
similarity feature
speech feature
