计算机工程与应用 ›› 2017, Vol. 53 ›› Issue (11): 178-181.DOI: 10.3778/j.issn.1002-8331.1512-0330

• 图形图像处理 • 上一篇    下一篇

基于八阶收敛牛顿迭代的Fast-ICA改进算法

陈  梦,何选森   

  1. 湖南大学 信息科学与工程学院,长沙 410082
  • 出版日期:2017-06-01 发布日期:2017-06-13

Improved Fast-ICA algorithm based on eighth-order convergence of Newton’s iterative method

CHEN Meng, HE Xuansen   

  1. College of Computer Science and Electronic Engineering, Hunan University, Changsha 410082, China
  • Online:2017-06-01 Published:2017-06-13

摘要: 解决盲源分离问题(BSS)最常用的方法是独立分量分析方法(ICA),快速独立分量分析方法(Fast-ICA)是目前广泛使用的独立分量分析方法。传统的Fast-ICA算法利用了二阶收敛的牛顿迭代方法进行优化,为了加快算法的收敛速度,提高算法的运行效率,利用八阶收敛的牛顿迭代方法对Fast-ICA算法进行优化,通过仿真验证了基于八阶收敛的Fast-ICA算法与传统的Fast-ICA和五阶收敛的Fast-ICA算法在分离性能上基本相同,但其具有更少的迭代次数和更快的收敛速率。

关键词: 盲源分离(BSS), 独立分量分析(ICA), 快速独立分量分析(Fast-ICA), 八阶收敛牛顿迭代

Abstract: The most popular solution for Blind Source Separation(BSS) problem is Independent Component Analysis(ICA) , and the Fast-ICA algorithm is widely used in BSS. The traditional Fast-ICA algorithm is optimized by the quadratic convergence of Newton iteration method. To accelerate the convergence speed and improve the running efficiency of the algorithm, this paper gives an improved Fast-ICA algorithm with eighth-order convergence of Newton iterative method. The simulation results show that the computational speed of the improved Fast-ICA is faster than that of the traditional Fast-ICA and the Fast-ICA with fifth-order convergence of Newton iteration method.

Key words: Blind Source Separation(BSS), Independent Component Analysis(ICA), Fast-Independent Component Analysis(Fast-ICA), eighth-order convergence of Newton iteration