Extraction of Fetal Electrocardiogram by Independent Component Analysis

摘要

The paper deals with a processing of the fetal electrocardiogram (fECG). The extraction of the fECG signal is made using Independent Component Analysis (ICA), which is a method very often used in the various areas of application. The accuracy verification is applied to the real data from the abdominal and direct fECG database. The Bland-Altman graph, was used for the evaluation of accuracy of the fetal heart rate variability (fHVR) estimation. It provides the number of the false positive values and false negative values enabling the calculation of the parameter of accuracy (ACC), sensitivity (SE), positive predictive value (PPV) and harmonic mean of the SE and PPV (F1). The proposed method reached the satisfying accuracy of the fHVR estimation at three recordings of total number 12. The ICA method has a predisposition to become a part of the hybrid fECG extraction systems.