Description: 摘要：提出了一种新的说话人识别中特征参数的提取方法*在分别使用傅立叶分析和小波分析得到两组特征
参数之后，进一步利用+,-./0准则进行参数选取，构造了一种新的混合特征参数*在不增加训练和识别时计算
量的同时，结合了傅立叶分析和小波分析两者的优点，具有更好的分类能力*实验结果显示，这种新的混合参数
有效地提高了说话人的识别率，能更好地表征说话人的特征*-Abstract: This paper presents a new kind of speaker recognition in the characteristic parameter extraction method*, respectively, using Fourier analysis and wavelet analysis of two sets of characteristic parameters, the further use of 2B !,-./ 0 criteria for parameter selection, construction A new mixed feature parameters* at no additional training and recognition computation At the same time, a combination of Fourier analysis and wavelet analysis of the advantages of both and has better classification capability* Experimental results show that the new mixing parameters effectively enhance the speaker Platform: |
Size: 139264 |
Author:王功臣 |
Hits:
Search - fisher criteria speaker recogniti