节点文献
应用小波包变换提取说话人识别的特征参数
Feature Extraction Using Wavelet Packet Transform in Speaker Recognition
【摘要】 论文研究了小波包变换,分析了MFCC参数的提取,比较了MEL滤波器组频带划分和小波包分解的频带划分,提取出了基于小波包变换的特征参数(DBWPTC),实验结果表明通过小波包变换提取的语音特征参数DBWPTC优于通过傅立叶变换提取的特征参数MFCC。
【Abstract】 This paper studies wavelet packet transform and analyzes the extraction of MFCC parameters.Comparing partition of frequency-band between MEL filter group and wavelet packet decomposition,we extract a new parameter,DBWPTC,based on wavelet packet transform.The experiment results indicate that the DBWPTC parameter is outperform MFCC.
【关键词】 说话人识别;
小波包变换;
MFCC;
矢量量化;
【Key words】 speaker recognition; wavelet transform; MFCC; vector quantization;
【Key words】 speaker recognition; wavelet transform; MFCC; vector quantization;
- 【文献出处】 计算机工程与应用 ,Computer Engineering and Applications , 编辑部邮箱 ,2006年09期
- 【分类号】TN912.34
- 【被引频次】30
- 【下载频次】418