Extraction Of Fetal Ecg Based On Ica

The extraction of fetal electrocardiogram (FECG) has much clinical significance in diagnosis of the fetal distress, oxygen deficit in womb and so on, it can reduce perinatal mortality. But the FECG acquired from the surface of mother body is very faint and often contains many noises and interferences. The strongest one is the maternal electrocardiogram (MECG) contributions, the maternal ECG and fetal ECG in time and frequency domain are mutual-aliasing, and they have randomness and stationary. At the same time, the FECG acquisition is highly sensitive to human interferences (maternal breathe, movement and muscle contraction), electrode placement and quantity, position of fetus, etc. So, how to obtain the clear fetal ECG. is still a uphill task.This paper first investigates research status of fetal ECG at home and abroad, researches the ICA application in the extraction of FECG, and aim to the FECG characteristics shows the extraction effectiveness by ICA.This paper researches the fast-point algorithm using negative entropy (Fast ICA), it can extracts fetal ECG fast and effectively. According to the nondeterminacy of separation results, adopts the spectrogram combined with time domain analysis to distinguish separated FECG. Simultaneously, for the inference and noise too much to separate FECG, adopts wavelet threshold denoising to preconditioning. Through the FECG extraction for analog signals and real signals, verifies the algorithm effectiveness. The simulation results show that after preconditioning, it can realize FECG effective segregation when low SNR, so it widens the application range.