论文部分内容阅读
该文提出一种结合小波变换(WPT)与快速独立分量分析(Fast ICA)算法的方法来分析脑电信号。首先,原始脑电信号是通过使用WPT分解为三个层。然后,设置第三层最高频率的系数为零,以减少脑电信号的随机噪声,同时尽可能的保留信号的细节。其次,采取快速独立分量分析算法的优势,从脑电信号中分离所有类型的噪声。提出一种准预期值(QEV)的方法确定脑电图信号来自何处。最后,为了检验系统的性能,所有信道的相关信号在快速独立分量分析的输出进行分析。实验结果证实,交叉相关系数是10-15或10-16的量级,几乎可以被视为零。所提出性能良好的方法可以去除脑电信号所有类型的噪声。
This paper presents a method combining wavelet transform (WPT) and fast independent component analysis (Fast ICA) algorithm to analyze EEG signals. First, the original EEG was decomposed into three layers by using WPT. Then, set the third highest frequency coefficient is zero, in order to reduce the random noise EEG signals, while retaining the details of the signal as much as possible. Second, take the advantage of a fast, independent component analysis algorithm that separates all types of noise from EEG signals. A quasi-expected value (QEV) method is proposed to determine where EEG signals come from. Finally, in order to test the performance of the system, all channel-related signals are analyzed at the output of the fast independent component analysis. Experimental results confirm that the cross-correlation coefficient is on the order of 10-15 or 10-16 and can be considered almost zero. The proposed method works well to remove all types of noise from EEG signals.