基于ICA的胎儿心电信号提取算法的比较

Comparison of FECG Extraction Algorithms Based on ICA

研究了2种独立分量分析(ICA)算法(扩展的信息极大算法和快速ICA算法)在胎儿心电信号提取中的应用.仿真实验采用混有母体心电和胎儿心电的混合信号,然后分别利用Fast-ICA算法和扩展Infomax算法提取胎儿心电信号.比较了2种算法的性能.仿真结果表明:在胎儿心电提取时,快速ICA算法更具有优越性,提取胎儿心电信号效果好,而且不需要进行学习速率的选择,收敛速度快,简单可靠.

Two algorithms of independent component analysis （ICA） are used to extract fetal ECG （FECG） in this article. Signals which mix the MECG signal and the FECG signal are used. In the simulation the two algorithms： Fast-ICA algorithm and Infomax algorithm, are checked and compared after they are used extract FECG signals. Conclusions can be drawn from the simulation that Fast-ICA is better in detecting the FECG and MECG, and it is fast convergent, simple and reliable because there is no need to select the learning rate.