摘要
在说话人识别系统中,语音特征参选是系统的关键问题之一.本文研究了MFCC参数、小波包分析.从听觉特性出发,提出基于小波包分析代替傅立叶变换的一种新的特征参数,给出了衡量各种特征参数识别能力的Fisher准则,结合Fisher准则构造一种新的混合特征参数,最后采用支持向量机实现说话人的分类识别.实验数据表明:有效地提高了说话人辨认系统的识别率.
In the speaker recognition system,the key problem is giving the personality feature of the speaker.This paper studied wavelet packet transform and analyzed the extraction of MFCC parameters.A new parameter was extracted using wavelet packet transform in place of Fourier transform.The Fisher criterion was used to measure the recognition of various parameters.A new hybrid parameter was constructed through a combination of the Fisher criterion.Finally,a support vector machine（SVM） was used to achieve the classification of Speaker Recognition.Experimental data show that the method is effective in raising the recognition rate of the speaker recognition system.
出处
《佳木斯大学学报（自然科学版）》
CAS
2010年第6期873-875,890,共4页
Journal of Jiamusi University：Natural Science Edition
基金
宿州学院校级自然科学研究项目(2009yzk23)
