采用声调基频特征的民族语声调识别  

Research on Tone Recognition of Chinese Minority Language Based on F0 Features

作　　者：龙润田 LONG Run-tian(School of Literature,Capital Normal University,Beijing 100089,China)

机构地区：[1]首都师范大学文学院,北京100089

出　　处：《贵州民族研究》2020年第11期123-127,共5页Guizhou Ethnic Studies

基　　金：国家社科基金青年项目“田野调查中的声调自动标注及校验方法研究”(项目编号:12CYY031)的阶段性成果。

摘　　要：为了实现对民族语无标注语料的声调自动识别,提高语言田野调查效率,语言学界提出了一种基于基频特征的声调聚类识别方法。在获得声调基频后,通过提取基频的曲线相关特征,采用聚类算法对提取的基频特征进行无监督聚类分析,从而获得声调识别的结果。识别测试表明,所述方法能较好地提高少数民族语言的声调识别准确性和精确度,在不同的少数民族语言的识别测试中均能够取得很好的效果。In order to realize the automatic recognition of the tone in the Chinese minority language,and the efficiency of the language filed investigation can be improved.This paper proposed tone recognition method based on tone features clustering.After extraction of acoustics characteristic parameters and fundamental frequency,it used method of linear curve fitting to extract the geometric features of F0.Finally the clustering algorithm is adopted to cluster analysis with tone features,the identification result can be obtained.Experiments on the minority languages and Chinese dialects tasks present the effectiveness and significant improvements in performance.

关 键 词：声调识别 民族语 基频特征 田野调查 

分 类 号：H2[语言文字—少数民族语言]