摘要
本文在盲信号处理的背景下介绍了独立分量分析算法的原理及特点。针对语音信号复杂多变、但源信号保持相互独立的特点,采用了改进后的FICA方法对混合语音信号进行分离,最终通过仿真实验,获得了满意的分离效果。
The principle and feature of independent component analysis （ICA） was illustrated as a branch of blind source signal processing. Due to the diversity of speech signal but which can keep only comparatively independence in the source signal, an enhanced method based on Fast ICA was used to separate the mixed speeches. As the final, a good result was obtained through emulating with Matlab.
出处
《微计算机信息》
2009年第28期176-177,96,共3页
Control & Automation
关键词
盲信号
独立分量分析
FICA
预处理
Blind source signal
Independent Component Analysis
Fast-ICA
Pre-processing
