Open AccessA Complete Adaptive Method for Fetal ECG Extraction Based on Single Channel
Author(s) -
Miao Zhang,
Guo Wei
Publication year - 2020
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/1621/1/012019
Subject(s) - pattern recognition (psychology) , artificial intelligence , computer science , hilbert–huang transform , feature extraction , filter (signal processing) , computer vision
Based on empirical mode decomposition (EMD) and coherent averaging method, a completely adaptive single channel fetal ECG (FECG) extraction method is proposed. EMD is used to decompose single-lead abdominal ECG (AECG) signals, so that the noise, FECG and MECG components main energy in AECG are distributed in different inherent mode functions (IMF), and R wave detection of MECG is performed by means of IMF, where MECG energy is dominant. MECG was extracted from AECG by means of coherent averaging based on R wave detection results. The remaining signal was named as noisy FECG. Usually, when MECG residue in noisy FECG is large, it needs to be further eliminated. To this end, the same method as above was used again to extract FECG from noisy FECG, and threshold method was used to eliminate MECG residue after extraction of FECG, and then superimposed onto FECG again, in order to recover the non-stationary feature lost in the coherent average process. The effectiveness of the proposed method is verified by the experiments of synthetic mixed signals and real database recorded signals.