英文发音中错误语音自动识别系统的设计.docVIP

英文发音中错误语音自动识别系统的设计 　　摘 要： 传统的英文发音识别系统对于学习者的错误发音不能及时进行反馈与纠正，存在误导学习者以及学习者英文水平提高缓慢的弊端。在此设计新的英文发音错误语音自动识别系统，其由语音录制模块、语音播放模块、英语发音评分模块和发音共振峰图像显示模块构成，给出评分模块的发音评分流程，实现英文发音的有效评分以及评分的存储，系统通过发音共振峰图形显示模块，清晰地表达出学习者发音与标准发音的不同之处，纠正其错误读音。通过英语音素检错程序使用独立阈值的方式来提高错误读音的检测性能，对不同音素用独立阈值进行衡量，使得英语发音中的错误语音自动识别结果更加科学化、精准化。实验结果表明，所设计的系统具有较高的错误语音自动识别能力。 　　关键词： 英语发音; 错误语音; 自动识别; 发音评分; 发音共振峰图像显示; 独立阈值 　　中图分类号： TN912.34?34 文献标识码： A 文章编号： 1004?373X（2018）10?0179?04 　　Abstract： The traditional English pronunciation recognition system fails to timely feed back and correct learners′ mispronunciation， which misleads learners and leads to learners′ slow improvement in English level. Therefore， a new error speech automatic recognition system for English pronunciation is designed. The system is composed of voice recording module， voice playing module， English pronunciation scoring module， and pronunciation formant image display module. The pronunciation scoring process of the scoring module is given to realize the effective scoring of English pronunciation and storage of scores. The pronunciation formant graphic display module is adopted to clearly express the differences between learners′ pronunciations and standard pronunciations， so as to correct their wrong pronunciations. The detection performance of wrong pronunciations is improved by using the independent threshold mode and the English phoneme error detection procedure. Different phonemes are measured with the independent threshold to make the wrong speech automatic recognition results of English pronunciation more scientific and accurate. The experimental results show that the designed system has a high error speech automatic recognition capability. 　　Keywords： English pronunciation; error speech; automatic recognition; pronunciation scoring; pronunciation formant graphic display; independent threshold 0 引 言 　　在?济全球化全面发展、我国对外开放进程不断推进的背景下，世界各国沟通往来不断加强，英语作为应用最频繁的语言，发挥了不可替代的作用[1]。在学习英语的过程中，存在学习者口语较差的