| 基于MFCC参数的说话人特征提取算法的改进 |
| |
| 引用本文： | 张晶,范明,冯文全,董金明.基于MFCC参数的说话人特征提取算法的改进[J].电声技术,2009,33(9):61-64,69. |
| |
| 作者姓名： | 张晶 范明 冯文全 董金明 |
| |
| 作者单位： | 北京航空航天大学,电子信息工程学院,北京,100191 |
| |
| 摘 要： | 在说话人识别系统中，特征参数的提取对语音训练和识别有着重要的影响。对于特征参数提取模块，提出了一种新的特征参数提取算法MFCC_E（Efficient MFCC）。相对于标准算法MFCC_S（StandardMFCC），MFCC_E在特征提取模块部分减少了53％的计算量。最终实验结果说明MFCC_E的识别率为90.3％，仅比标准MFCC算法92．0％的识别率降低1．7％。因为MFCC_E算法的这种特点，使其能够更有效的适用于硬件实现。
|
| 关 键 词： | 特征提取 MFCC_S MFCC_E |
An Efficient Speaker Feature Extraction Method Based on MFCC |
| |
| ZHANG Jing,FAN Ming,FENG Wen-quan,DONG Jin-ming.An Efficient Speaker Feature Extraction Method Based on MFCC[J].Audio Engineering,2009,33(9):61-64,69. |
| |
| Authors: | ZHANG Jing FAN Ming FENG Wen-quan DONG Jin-ming |
| |
| Affiliation: | ZHANG Jing,FAN Ming,FENG Wen-quan,DONG Jin-ming (School of Electronic and Information Engineering,Beihang University,Beijing 100191,China) |
| |
| Abstract: | Feature extraction is a significant module for speech training and recognition in speech recognition system. A new algorithm of feature extraction MFCC_E(Efficient MFCC) is introduced. Compared to the standard algorithm MFCC_S (Standard MFCC),the new algorithm reduces the computation power by 53%. The simulation results indicate MFCC_E has a recognition accuracy of 90.3%,and there is only an 1.7% reduction compared to MFCC_S which has 92.0% recognition accuracy. The new algorithm is acceptable for hardware ... |
| |
| Keywords: | MFCC_S MFCC_E |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
