Fetal electrocardiogram extraction using Robust Independent Component Analysis

The extraction of Fetal Electrocardiogram (FECG) from maternal cutaneous electrocardiogram (ECG) recordings is of prime importance in assessing fetal well being for antenatal and postnatal care. However, these recordings are contaminated with various kinds of noise like power-line interference, base line wander, high frequency noise, electromyogram (EMG) noise etc. This paper proposes a frequency domain approach to the extraction of FECG using Robust Independent Component Analysis (RobustICA). The efficiency of the proposed method for FECG extraction and artifact removal is validated using real and synthetic ECG signals.