一种基于二代小波变换与盲信号分离的脑电信号处理方法  

A Processing Method of EEG Signals Based on Second Generation Wavelet Transform and Blind Signal Separation

作　　者：罗志增[1] 李亚飞[1] 孟明[1] 孙曜[1] 

机构地区：[1]杭州电子科技大学机器人研究所,浙江杭州310018

出　　处：《航天医学与医学工程》2010年第2期137-140,共4页Space Medicine & Medical Engineering

基　　金：国家863项目(2008AA04Z212);国家自然科学基金(60874102;60705010)

摘　　要：目的研究对混杂有眼电和心电干扰脑电信号的处理方法。方法首先用二代小波硬/软阈值、折衷阈值、μ律阈值方法对脑电信号消噪,然后运用FastICA算法对消噪后仍含眼电和心电的脑电信号进行盲信号分离。结果二代小波μ律阈值方法对脑电信号有较好的消噪效果,FastICA算法能成功分离出脑电中眼电和心电的干扰。结论运用二代小波μ律阈值法对脑电消噪后再用FastICA算法对独立源产生的干扰进行分离是一种有效的预处理方法。Objective To study a processing method for EEG signals mixed with EOG and ECG signals disturbance.Methods First,the EEG was denoised by the hard threshold method,the soft threshold method,the compromise threshold method and the μ law threshold method in the second generation wavelet,and then the denoised EEG which still contained EOG and ECG was separated by fast independent component analysis（ FastICA） algorithm.Results The μ law threshold method of the second generation wavelet had better denoising effect and FastICA algorithm had more ideal separate performance.Conclusion It is an effective preprocessing method for EEG in denoising with the μ law threshold method of the second generation wavelet and then in separating disturbance of independent source with FastICA algorithm.

关 键 词：脑电信号 二代小波 μ律阈值法 消噪 FASTICA算法 

分 类 号：R318.04[医药卫生—生物医学工程] TP391[医药卫生—基础医学]