Joint blind source separation algorithms in the separation of non-invasive maternal and fetal ECG

Blind source separation (BSS) is used in many fields of signal and image processing. It is used for separating a set of source signals from mixed signals without the aid of information about the source signals or the mixing process. The paper mainly focuses on (1) Separation of maternal and fetal ECG signal (2) Performance measure of various BSS and JBSS algorithm in term of SIR and execution time. Various BSS algorithm like ICA, FASTICA, JADE and JBSS algorithms like MCCA, SOBI, JBSS_SOS, JBSS_CUM4 are used for blindly separating the source signals. And the impacts on separation of maternal and fetal signal are examined. The simulations are conducted in MATLAB using Non-Invasive ECG of pregnant women from PhysioNet database. The JBSS_CUM4 algorithm shows better performance with the SIR value of 29.58dB for MECG and 7.44dB for FECG.