| 基于小波包分解的抗噪说话人识别特征参数 |
| |
| 引用本文： | 张昊慧.基于小波包分解的抗噪说话人识别特征参数[J].通信技术,2010,43(12):144-146. |
| |
| 作者姓名： | 张昊慧 |
| |
| 作者单位： | [1]准阴师范学院物理与电子电气工程学院,江苏淮安223001 [2]东南大学信息科学与工程学院,江苏南京210096 |
| |
| 摘 要： | 为了提高说话人识别中语音特征参数的鲁棒性,提取了新的特征参数DWT-MFCC,在提取该参数时利用了新构造的阈值函数,并基于高斯混合模型的说话人识别系统进行实验。实验结果表明,相对于传统的MEL倒谱系数（MFCC）参数,在相同的噪声环境下,DWT-MFCC参数具有更高的说话人识别率。
|
| 关 键 词： | 小波包变换 MFCC参数 阈值函数 GMM系统 |
Characteristic Parameter based on Wavelet Packet Decomposition in Speaker Recognition |
| |
| ZHANG Hao-hui.Characteristic Parameter based on Wavelet Packet Decomposition in Speaker Recognition[J].Communications Technology,2010,43(12):144-146. |
| |
| Authors: | ZHANG Hao-hui |
| |
| Affiliation: | ZHANG Hao-hui(①School of Physics and Electronic Electrical Engineering,Huaiyin Normal University,Huai＇an Jiangsu 223001,China; ②School of Information Science and Engineering,Southeast University,Nanjing Jiangsu 210096,China) |
| |
| Abstract: | To improve the performance of speaker recognition in noisy environment,a new parameter,DWT-MFCC,is extracted.Based on a new threshold function and the GMM system the speaker recognition is evaluated.The results show that,as compared with the MFCC parameter,the DWT-MFCC has higher recognition accuracy in noisy environment. |
| |
| Keywords: | wavelet packet transform MFCC parameter threshold function GMM system |
| 本文献已被 维普 万方数据 等数据库收录！ |
|
