摘要
在实验室录音设备条件下,提取基于Mel的特征参数及其差分倒谱系数(DCC)和差分频谱倒谱系数(DSCC),利用增减分量法进行有效融合,建立基于VQ的说话人识别模型,采用LBG算法设计不同码本容量进行说话人识别实验,并在各类噪声环境下验证混合参数性能.实验结果表明,与单一参数及其他融合参数相比该混合特征参数具有更高的识别率和鲁棒性.
With the recording voice library in the laboratory, we extract Mel frequency cepstral coefficient, its deltacepstral coefficient and delta-spectral eepstral coefficient, and fuse them with change component method effectively. And a speaker recognition model is established based on VQ model using LGB algorithm to design different codebook capacity for the experiment and we evaluate the hybrid feature parameters in noise environments. Experimental results show that by applying the fusion hybrid feature parameters the recognition system rate and robustness are obviously improved compared with other parameters.
出处
《微电子学与计算机》
CSCD
北大核心
2016年第4期60-63,68,共5页
Microelectronics & Computer
基金
国家自然科学基金(61363005
61462017)
广西研究生教育创新计划资助项目(YCSZ2015152)
关键词
说话人识别
VQ
混合参数
鲁棒性
speaker recognition
VQ
hybrid feature parameters
robustness
