数字滤波和小波变换提取脑电节律的比较  

Comparison of Digital Filter and Wavelet Transform for Extracting Electroencephalogram Rhythm

作　　者：谢涛嵘[1] 裴剑[1] 贾彩丽[2] 陈树德[1] 乔登江[1] 

机构地区：[1]华东师范大学物理系精密光谱科学与技术国家重点实验室 [2]上海电力学院

出　　处：《生物医学工程学杂志》2009年第4期743-747,共5页Journal of Biomedical Engineering

基　　金：长江学者和创新团队发展计划项目资助;上海市重点学科建设项目资助(B408);上海高校选拔培养优秀青年教师科研专项基金项目(Z-2006-90)

摘　　要：提取脑电节律往往是临床脑电图(EEG)诊断的关键一环。根据小波多分辨滤波器组的特点,从理论和数值实验两方面来讨论小波分解对信号频段的二进划分的特点,把一段采样频率为2kHz的EEG信号用4阶Daubechies小波(db4)作8层分解,重构各个分量,得到它的δ、θ、α、β、γ节律。由于小波多分辨率滤波器组具有品质因数恒定的特性,使得该方法能从宽的频谱范围中提取出低频段的脑电节律。和用6阶椭圆无限长冲激响应(IIR)数字滤波器滤波的效果相比,有其显著的优越性。It is very important to extract electroencephalogram （EEG） rhythm in clinical diagnoses. Digital filter and wavelet transform are used to extract the rhythm from a piece of EEG at the sampling rate of 2 kHz. The Dauhechies order 4 wavelet （db4） was used to decompose the EEG at 8 levels. According to the filter characteristic of wavelet decomposition, the reconstructions of a8, d8, d7, d6 and d5 component are nearly corresponding to the rhythms of δ、θ、α、β、γ separately. The 6 order ellipse infinite impulse response （IIR） filter is also used to decompose the EEG. As the quality factor of wavelet decomposition filter is constant, the wavelet transform obtains better extracted rhythm than the digital filter. Furthermore, the wavelet transform method can be used to extract the low frequency rhythm from wide frequency band.

关 键 词：脑电图 小波分解 品质因数 椭圆滤波器 

分 类 号：R318.04[医药卫生—生物医学工程] O177[医药卫生—基础医学]