A New Fetal Electrocardiogram Extraction Algorithm

This paper proposes an algorithm for Fetal Electrocardiogram (FECG) extraction, which is based on joint maximization of kurtosis and autocorrelation. It fully exploits the FECG's special temporal property and its independence on other sources (such as some noise, the maternal electrocardiogram); therefore it is able to extract clearer FECGs, compared with existing algorithms. Extensive experiments show that the proposed algorithm can rapidly and efficiently extract FECGs and greatly reduce lots of noise.