| 快速定点独立向量分析在语音信号盲分离中的应用 |
| |
| 引用本文： | 黄会营,李小光,洪俊芳.快速定点独立向量分析在语音信号盲分离中的应用[J].河南师范大学学报(自然科学版),2012,40(6):147-149,164. |
| |
| 作者姓名： | 黄会营 李小光 洪俊芳 |
| |
| 作者单位： | 1. 洛阳理工学院电气工程与自动化系,河南洛阳,471003
2. 中国人民解放军96544部队,河南洛阳,471003 |
| |
| 基金项目： | 河南省科技厅自然科学基金，洛阳理工学院青年基金 |
| |
| 摘 要： | 针对包括环境噪声的混叠语音信号分离困难问题,提出一种新的频域快速定点独立向量分析模型,对混叠语音信号进行盲分离.该模型中的每个源信号以向量形式混合,模型采用二次泰勒多项式把牛顿迭代算法直接应用于复变量对照函数,通过最优化使牛顿迭代规则更加简单和清晰.对实际采集的语音信号进行仿真,实验证明算法很好地分离了卷积盲源信号,新算法具有良好的收敛性和较高的准确性.
|
| 关 键 词： | 独立向量分析 卷积混合 盲源分离 快速算法 |
Fast Fixed-point Independent Vector Analysis Algorithms for Speech Blind Separation |
| |
| HUANG Hui-ying
,
LI Xiao-guang
,
HONG Jun-fang.Fast Fixed-point Independent Vector Analysis Algorithms for Speech Blind Separation[J].Journal of Henan Normal University(Natural Science),2012,40(6):147-149,164. |
| |
| Authors: | HUANG Hui-ying LI Xiao-guang HONG Jun-fang |
| |
| Institution: | 1.Department of Electrical Engineering and Automation,Luoyang Institute of Science and Technology,Luoyang 471003,China; 2.Chinese People’s Liberation Army 96544 Troop,Luoyang 471003,China) |
| |
| Abstract: | In order to solve the difficulty of separating for the aliasing voice signals including ambient noise.A new model,called independent vector analysis,is proposed to blind separation for aliasing speech signal.Each of the source signals are mixed component-wise,Newton iterative algorithm is applied directly to a contrast function of complex variables by using quadratic Taylor polynomials in the model.The Newton derivation rules become more simple and clear by optimization.The actual acquisition of the voice signal is used to simulate based on the new model.The experimental results show that the algorithm has good convergency and high accuracy. |
| |
| Keywords: | independent vector analysis convolutive mixture blind source separation fast algorithm |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
|
