基于GMM-UBM模型的语言辨识研究

Automatic language identification based on GMM-UBM

与说话人识别、连续语音识别相比,自动语言辨识是一个相对较新的研究,而且是一项较难的课题。本文给出了一种基于GMM-UBM模型的语言辨识系统,并利用OGI-TS电话语音库对算法的性能进行了测试,然后给出了实验结果。实验结果表明,该算法也是进行语言辨识的一种有效方法。

Compared with other speech technologies in speech processing, automatic language identification is a relatively new yet difficult problem. In this paper, a language identification algorithm is provided and some experiments are conducted using OGI-TS telephone speech corpus. Then experiments results are described. It is shown that GMM-UBM is another efficient method to language identification problems.