一种改进的VQ算法在说话人识别上的应用

Application of the improved VQ arithmetic to speaker recognition

本文介绍了一种基于动态权值改进的矢量量化(VQ)方法。根据不同环境下不同说话人语音特征各阶差异较大的特点.对矢量量化算法进行了改进。经实验结果证明.该方法提高了说话人识别系统的识别率。

In this paper, we introduce an improved VQ arithmetic. This method based on the different character of the speech in different environment and special speaker. It improve the arithmetic of vector quantization. Through a series of experiments , show that the new VQ arithmetic, reform the performance of the system, boost the rate of the speaker recognition.