摘要
提出了一种以元音的共振峰模式为特征基于支持向量机算法的分类评估方法,用以对汉语普通话中的韵母发音水平进行客观测试。此算法为每个韵母分别训练全分类模型、子分类模型和评估模型,在两级分类的基础上对发音水平进行测试打分。实验结果表明,全分类模型可以达到90%以上的分类正确率,客观测试与专家主观评估的相似度达到82%,在性能上超过了传统的以倒谱系数为特征的隐含马尔科夫模型方法。
A method used for objective evaluation of pronunciation of vowels in standard Chinese is presented. The formant patterns of vowels are selected as the main features and an improved evaluation algorithm based on Support Vector Machine is proposed. In this algorithm, two-level classification strategy is employed. A full-classification model and a sub-classification model are trained for each vowel. The pronunciation quality is evaluated based on the classification results of this two-level strategy with evaluation model of each vowel. The new evaluation method is compared with traditional methods such as HMM posterior probability scoring method and feature of Mel-Frequency Cepstrum Coefficients （MFCC）, and the results show that the performance is effectively improved by the proposed method. The correlation of scores between human testers and machine has achieved 82%.
出处
《声学学报》
EI
CSCD
北大核心
2007年第2期122-128,共7页
Acta Acustica
基金
国家973项目支持(2004CB318106)
