基于小波奇异点检测和阈值去噪的眨眼伪迹去除方法

目的眨眼伪迹是脑电中一种常见且影响严重的伪迹。本论文提出一种基于小波奇异点检测和阈值去噪的眨眼伪迹去除方法,无需眼电参考信号,做到自动去除单导脑电信号中的眨眼伪迹。方法首先利用小波奇异点检测特性以检测眨眼伪迹的峰值位置,然后只对眨眼伪迹区域进行小波阈值去噪。结果实验结果表明,本方法能够有效检测眨眼伪迹,避免了普通方法去噪时对非眨眼区域的影响。结论本方法使用的阈值和阈值函数优于典型的阈值和软、硬阈值函数,有效地去除了脑电中的眨眼伪迹。

A method for blink artifact removal based on wavelet singularity detection and thresholding denoising

Objective Blink artifact is common and has serious effect on EEG .To remove the blink artifacts in EEG automatically without a reference channel , this paper proposes a method for blink artifact removal based on wavelet singularity detection and thresholding denoising .Methods First, the detection property of wavelet singularity was used to detect the positions of blink artifact peaks , and then only the blink artifact zones were denoised by the wavelet thresholding method .Results The experimental results showed that the proposed method could effectively detect the blink artifacts and avoid affecting the EEG outside the blink artifact zones in usual methods .Conclusions The threshold and thresholding function used in the paper could effectively remove the blink artifacts in EEG and outperform the conventional soft or hard thresholding estimators.