独立分量分析及其在生物医学工程中的应用

Independent component analysis and its BME application

:独立分量分析 ( Independent Component Analysis,简记 ICA)是信号分解技术的新发展。ICA与 PCA(主分量分析 )或 SVD(奇异值分解 )的主要不同是 :后者分解得的各分量只是互不相关 ,而前者则要求各分量相互统计独立。体表测量得的信号往往包含若干相对独立的成分 ,因此采用ICA技术来分解 ,所得结果往往更有生理意义 ,有利于去除干扰和伪迹。本文简短地回顾 ICA的基本原理、判据、算法和其在生物医学工程中的应用 ,并作出展望及指出存在问题。

Independent Component Analysis(ICA)is a new development of signal decomposition.The main difference between ICA and PCA(Principal Component Analysis)or SVD(Singular Value Decomposition) is that the components decomposed by the later method are only mutually uncorrelated,whereas the components decomposed by the former method are mutually independent statisically.Since the signals measured on the surface of human body are usually the mixture of several relatively independent sources,their ICA decomposition can usually lead to results more plausible physiologically.A short review on the basic principles,criteria and algorithms of ICA is given in this paper,together with some examples of its BME application.Perspectives and open questions are also addressed.