| 基于FastICA的混合音频信号盲分离 |
| |
| 引用本文： | 贾银洁,许鹏飞. 基于FastICA的混合音频信号盲分离[J]. 信息与电子工程, 2009, 7(4): 321-325 |
| |
| 作者姓名： | 贾银洁 许鹏飞 |
| |
| 作者单位： | 1. 淮阴工学院,计算机工程系,江苏,淮安,223001
2. 西安电子科技大学,ISN国家重点实验室,陕西,西安,710071 |
| |
| 摘 要： | 独立成分分析（ICA）作为一种有效的盲源分离技术已成为信号处理领域的热点，它以非高斯源信号为研究对象，在统计独立的假设下，对多路观测到的混合信号进行盲信号分离。为了提高算法的收敛速度和稳态精度，介绍了独立成分分析的基本原理，以及利用FastICA算法进行信号分离的理论依据，引入了改进的非线性函数，运用Matlab进行仿真比较3种非线性函数下的分离性能和改进的非线性函数在不同θ下的分离性能，结果表明在综合因素的考虑下，该改进函数在实现混合音频信号盲分离方面比改进前更有效。
|
| 关 键 词： | 独立成分分析 音频信号 FastICA算法 负熵 |
Blind separation of mixed audio signal based on FastICA |
| |
| Affiliation: | JIA Yin-jie, XU Peng-fei (1.Department of Computer Engineering, Huaiyin Institution of Technology, Huaian Jiangshu 223001, China; 2.State Key Laboratory of Integrated Service Networks, Xidian University, Xi＇an Shaanxi 710071, China) |
| |
| Abstract: | The independent component analysis as a method widely used in blind source separation is a hotspot in signal processing. It addresses non-gaussian source signals under assuming independent of each other, it performs blind separation for muhichannel observed signals. In order to improve the convergence speed and precision in steady state, the basic principle of the independent component analysis and the theory for signal separation by the FastICA were presented. In order to improve the performance of the separation, the improved nonlinear function was introduced. The separating performance of three nonlinear functions and the improved nonlinear function under different parameter 0 were compared by simulation using Matlab. The results showed that based on an overall consideration, the improved function was more effective than the former in realizing the blind separation of mixed audio signals. |
| |
| Keywords: | Independent Component Analysis（ICA） audio signal FastlCA negentropy |
| 本文献已被 维普 万方数据 等数据库收录！ |
|
