TWIN FETUS ECG SIGNAL EXTRACTION BASED ON TEMPORAL PREDICTABILITY

Abstract

The most popular defects that infect new born babies have close connection with heart. Approximately 1% percent of new born babies suffering from defects that are caused by heart. By observing Electrocardiogram "ECG" during gestation doctors can study the fetal heart activity that collected from mother abdominal and rectify number of defects which has been diagnosed in fetal heart. Many techniques like filters, BSS and even artificial intelligence used to extract and process fetal ECG signal. In case of twins gestation the problem would be more complicated because the recorded signal is a mixture of multi signals which are mother ECG, fetuses ECG and noises where every signal comes from different source in addition fetus have the same ECG signal features as well. In this study the concentration would be on BSS techniques specifically on Stone BSS method. A comparison has been made between Stone method and two other methods "EFICA and JADE". The results proved that Stone method has better performance comparing to the other BSS techniques.