优先度排序RBF神经网络在与文本无关说话人确认中的应用

邓浩江; 王守觉; 杜利民

优先度排序RBF神经网络在与文本无关说话人确认中的应用

邓浩江^{①;王守觉},
王守觉,
杜利民

计量
- 文章访问数: 2383
- HTML全文浏览量: 107
- PDF下载量: 467
- 被引次数: 0
出版历程
- 收稿日期: 2002-04-11
- 修回日期: 2002-09-16
- 刊出日期: 2003-09-19

Text-independent speaker verification using priority ordered radial basis function networks

Deng Haojiang^{①;王守觉},
Wang Shoujue,
Du Limin

摘要

摘要: 该文介绍了优先度排序径向基函数(PORBF)神经网络的结构与算法,并提出了将其应用于与文本无关说话人确认时的训练算法、似然度的计算方法以及识别规则。为了增强PORBF网络的泛化能力,该文用压缩矢量构造抑制样本集,提出了顺序选取、最近邻选取和最远距离选取等3种选择抑制样本集中说话人的方法,并对PORBF神经元的输出进行了等比递减加权.在相同条件下的与文本无关说话人确认实验中,传统的矢量量化方法的等差错率可达10.56％,而基于PORBF网络的确认系统使用最近邻选择方法构造抑制样本集,其等差错率可达6.83％;性能提高很多。
- 优先度排序; 说话人确认; 与文本无关; RBF网络
Abstract: The structure and algorithm of Priority Ordered Radial Basis Function (POR.BF) Networks is introduced. The concrete training algorithm, calculational methods of likelihood score and verification rule, used for text-independent speaker verification, are proposed. To en-hance the generalization ability, the compressing vectors are applied to construct the inhibitory samples set and three methods including sequential selection, nearest neighbor selection and furthest distance selection are presented for the choose of anti-speakers. Moreover, the out-puts of neurons are weighted by a descendent array. Using these algorithms and methods, the performance is examined by a series of experiments. The results show that under the identi-cal experiment conditions, when the inhibitory set is composed of anti-speakers compressing vectors selected using nearest neighbor method, the Equal Error Rate (EER) using PORBF networks can decreased to 6.83% from 10,56% using conventional VQ method. For speaker verification, the PORBF network provides better performance than the VQ classifier.

HTML全文

参考文献(1)

K R.Farrell,S.Kosonocky, R.J.Mammone,Neuraltree network/vector quantization probability estimator for speaker recognition,Proc.IEEE,1994,82(1),279-288.[2]Zhang Yiying,Zhang David,Zhu Xiaoyan,A novel text-independent speaker verification method based on global speaker model,IEEE Trans.on System,Man,and Cybernetics,2000,30(5),598-602.[3]史静朴,等,用神经计算机的说话人确认系统及其应用,电子学报,1999,27(10),1999,27-29.[4]R. A. Sukkar, M. B. Gandhi, A. R. Setlur, Speaker verification using mixture decomposition discrimination, IEEE Trans. on Speech and Audio Processing, 2000, 8(2), 292-299.[5]K.R. Farrell, R. J. Mammone, K. T.assaleh, Speaker recognition using neural networks and conventional classifiers, IEEE Trans. on Speech and Audio Processing, 1994, 2(1), 194-204.[6]J. Oglesby, J. S. Mason, Radial basis function networks for speaker recognition, Proc. ICASSP,New Mexico, USA, 1990, 393-396.[7]R.A. Finan, A. T. Sapeluk, Text-independent speaker verification using predictive neural networks, IEE Conf. Pub. of ANN, Cambridge, UK, July, 1997, 274-279.[8]D.S. Broomhead, D. Lowe, Multivariable functional interpolation and adaptive networks, Complex Systems, 1988, 2(2), 321-355.[9]J. Moody, C. J. Darken, Faster learning in networks of locally-tuned processing units, Neural Computation, 1989, 1(2), 281-293.[10]Wang Shoujue, Priority ordered neural networks with better similarity to human knowledge representation, Chinese Journal of Electronics, 1999, 8(1), 1-4.[11]陈川,方向基函数神经网络模型与算法及其在模式识别中的应用,[博士论文],北京,中国科学院半导体研究所,1999.[12]Liu Chi-Shi, Wang Hsiao-Chuan, Lee Chin-Hui, Speaker verification using normalized loglikelihood score, IEEE Trans. on Speech and Audio Processing, 1996, 4(1), 56-60.[13]Y. Linde, A. Buzo, R. M. Gray, An algorithm for vector quantizer design, IEEE Trans. on Commun., 1993, COM-28(1), 84-95.

施引文献

资源附件(0)

访问统计