基于MMTD与小波硬阈值的脑电信号去噪方法

De-noising Method of EEG Signal Based on MMTD and Wavelet Hard-threshold

针对小波硬阈值去噪算法在脑电信号去噪过程中会导致部分有效信号丢失的不足,提出基于中介真值程度度量(Measuring of Medium Truth Degree,MMTD)与小波硬阈值相结合的脑电信号去噪(MMTD and Wavelet Hard-threshold,MAWH)方法。该算法的基本思想是采用小波变换对含噪信号进行分解,对分解后的各层高频小波系数进行阈值处理,对处理后的小波系数重构,以达到消噪的目的。通过仿真试验,采用经典的RMSE、SNR评价标准,将MAWH方法与典型的硬、软两种阈值法进行比较。结果表明,在不同的噪声强度下,MAWH去噪方法对脑电信号的去噪效果更加优越。

To overcome the shortage of losing partial important information of hard-threshold method with EEG signal de-noising process, a novel de-noising method based on the combination of measuring of medium truth degree （MMTD） and EEG is proposed. By decomposing noisy signals of wavelet transform, handling threshold of high-frequency wavelet coefficients in every layer, and reconstructing post-processing of the wavelet coefficients, the purpose of noise elimination can be guaranteed. Under different noise intensity, the experimental results show that the MAWH （MMTD and wavelet hard-threshold） method has perspective of lower RMSE and higher SNR compared to hard-threshold and soft-threshold.