脑电棘波识别和噪声消除的小波变换方法

EEG spikes detection and denoising methods based on wavelet transform

研究了利用二进小波变换的模极大值识别脑电信号奇异点如棘波和消除噪声的方法,该方法在较好保留原脑电信号奇异信息的同时能有效地消除噪声。进一步讨论了信号与白噪声的奇异性指数的区别,以及小波变换模极大值沿各变换尺度传递的不同特性,并利用该特性区分信号中的奇异点和噪声,能准确识别奇异点的位置。这种奇异性识别技术在信号的特征提取和消除噪声方面有广阔的应用前景。

In this paper,the EEG signal singularity detection and denoising methods based on the dyadic wavelet transform(WT)modulus maxima are studied,and the denoising method can remove noise effectively as well as keep original EES singularity.The singularity Lipschitz exponent difference between signal and noise is discussed.We also compare the performance of signal and noise's WT at difference scale.The EES singularity points(spikes)locate in the positions where the modulus maxima match at different scales.This methods may be applied to signal characteristic extraction and denoising.