EMD-BASED ECG DENOISING USING SOURCE SEPARATION
EMD-BASED ECG DENOISING USING SOURCE SEPARATION
复制标题
DOI：
10.1142/s0219519415500827
复制
发表时间：
2015-10
影响因子：
0.8
通讯作者：
Mourad Talbi;S. Abid;A. Cherif
中科院分区：
文献类型：
--
作者：
Mourad Talbi;S. Abid;A. Cherif
我们考虑使用源分离的心电图（ECG）去噪问题。在本研究中，提出了一种使用经验模式分解（EMD）和源分离的混合技术。该技术由两个步骤组成，第一步是将 EMD 应用于两种不同的混合物。这些混合物是通过以加性方式用具有两个不同信噪比 (SNR) 值的高斯白噪声破坏相同的 ECG 信号而获得的。第二步是计算每个获得的本征模态函数 (IMF) 的熵，并找到具有最小熵的两个 IMF。这两个 IMF 用于通过源分离来估计 ECG 信号与噪声的分离矩阵。通过将所提出的技术与基于使用 Runica 的时域源分离的去噪技术和基于仿生小波变换 (BWT) 并使用源分离的技术进行比较来进行评估。从 SNR 和均方误差 (MSE) 计算获得的结果表明，所提出的技术优于本次评估中使用的其他两种技术。
We consider the problem of electrocardiogram (ECG) denoising using source separation. In this study, a hybrid technique using empirical mode decomposition (EMD) and source separation, is proposed. This technique consists of two steps, the first step consists in applying the EMD to two different mixtures. These mixtures are obtained by corrupting in additive manner, the same ECG signal by a white Gaussian noise with two different values of the signal-to-noise ratio (SNR). The second step consists in computing the entropy of each obtained intrinsic mode function (IMF) and finds the two IMFs having the minimal entropy. These two IMFs are used to estimate the separation matrix of the ECG signal from noise by using the source separation. The proposed technique is evaluated by comparing it to the denoising technique based on source separation in time domain using runica and the technique based on Bionic wavelet transform (BWT) and also using source separation. The obtained results from SNR and the mean square error (MSE) computations, show that the proposed technique outperforms the other two techniques used in this evaluation.