期刊文献+

基于GMM与改进MCE训练的说话人识别研究

Study on Speaker Recognition Based on GMM and Improved MCE
下载PDF
导出
摘要 在应用最小分类错误(MCE)准则对识别说话人的高斯混合模型(GMM)训练中,采用一个权重函数来确定说话人模型参数调整量的权值,使得比较近的竞争说话人模型的权值大,比较远的竞争说话人模型的权值小。并采用梯度概率递减算法来实现损失函数的最小化,有效提高了说话人识别的速度和精度。 In the paper,Minimum Classification Error(MCE) criterion is used to train Gaussian Mixture Model(GMM) for Speaker Recognition.A weight function is determined to adjust the weight of speakers to assign a greater weight for more similar speaker.Algorithm of gradient probability degression is utilized to minimize lost function.The presented method has improved speed and accuracy of speaker recognization.
出处 《自动化与仪器仪表》 2010年第6期21-23,共3页 Automation & Instrumentation
关键词 说话人识别 高斯混合模型 最小分类错误 Speaker Recognition Gaussian Mixture Model Minimum Classification Error
  • 相关文献

参考文献1

二级参考文献5

  • 1Mammone R J. , Xiaoyu Zhang. Ramachandran R. P. ,Robust speaker recognition: A feature--based approach [J]. IEEE Signal Processing Magazine, 1996,13 (5) :58-71.
  • 2Rong Zheng, Shuwu Zhang, Bo Xu. Text-independent speaker identification using GMM-UBM and frame level likelihood normalization[J]. Chinese Spoken Language Processing, 2004 International Symposium on, 2004,12 :289 -292.
  • 3L. E. Baum, T. Petrie, G. Soules, et al. A maximization technique occurring in the statistical analysis of probabilistic function of markov model[M]. ANN. math. Statist, 1970.
  • 4Cardoso J F. Infomax and maximum likelihood for source separation[J]. IEEE Letters on Signal Processing, 1997,4(5) : 112- 114.
  • 5邵央,刘丙哲,李宗葛.基于MFCC和加权矢量量化的说话人识别系统[J].计算机工程与应用,2002,38(5):127-128. 被引量:34

共引文献2

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部