对偶树复小波在脑电消噪中的应用

Application of dual-tree complex wavelet transform in EEG denoising

为解决脑电去噪过程中离散小波带来的信息丢失与频率混叠问题,提出了一种新型对偶树复小波去噪方法.用对偶树复小波对输入脑电信号(EEG)进行多层分解,得到实树部分与虚树部分,分别对实树部与虚树部各子代小波系数进行小波中值阈值处理,再逆变换得到去噪小波.仿真结果表明:该方法可以比传统离散小波去噪方法获得更好的信噪比与均方误差,因此更适合于处理微弱的脑电信号.

To solve information loss and frequency aliasing by discrete wavelet transform in the process of electroencephalogram (EEG)denoising,a new EEG signal denoising algorithm was pro-posed,which was based on dual-tree complex wavelet transform.The dual-tree complex wavelet transform was used to conduct a multilayered decomposition on the EEG inputted,so the real tree and the imaginary tree could be obtained,and the median threshold function was used to process the off-spring wavelet coefficients of the real tree and the imaginary tree,then the denoised wavelet was ob-tained by the method of inverse transformation.Simulation results reveal that the SNR and mean square error (MSE)of the proposed method are better than those of traditional discrete wavelet de-noising method,and the proposed method is more suitable for processing the weak EEG signal.