基于Multilingual的音素识别及其在语种识别中的应用

Multilingual-based PRLM for language identification

国际上对自动语种识别进行了广泛的研究,提出了各种各样的方法,美国国家标准技术研究所(NIST)多年的评测表明,基于并行音素识别(parallel phoneme recognition language modeling,PPRLM)的方法取得了很好的性能。该文提出了一种基于多种语言的音素识别方法的自动语种识别系统,系统中Multilingual音素集是使用基于数据驱动聚类获得。通过真实环境电话语音测试表明,该方法在只使用了很少的识别时间的情况下,获得了跟传统的PPRLM系统可比的识别正确率。同时经过与PPRLM系统融合后,获得了更好的性能,跟其他主流的几种语种识别方法也有可比的性能。

A variety of methods have been developed for automatic language identification.Evaluations by National Institute of Standard and Technology(NIST) how that the parallel phoneme recognition language modeling(PPRLM) algorithm gives good results.This paper presents a multilingual-based phoneme recognition language modeling(PRLM) system for automatic language identification.The multilingual phone-set was obtained by clustering data-driven language-dependent phone sets.Evaluations of telephone speech in real envi...