Application of SVM Algorithm for Fetal ECG Extraction from a Single Maternal Abdominal Record

Abstract

The electrical activity of the foetal cardiac muscles is known as foetal ECG (FECG), and it can provide crucial details on the health of the fetus's heart. A pregnant woman's belly can be used to non-invasively capture this signal during pregnancy. However, since additional sources of noise, including the maternal ECG generally overpower the FECG recording, it would be ineffective. However, a clean FECG may be retrieved from the abdominal recording if it is correctly processed, and FECG can be used to evaluate the functioning of the foetal heart. In order to extract foetal electrocardiogram (ECG) from a single abdomen record, a unique two-tier approach is presented in this work. The abdominal signal is processed through a smoothing filter in the first layer of the proposed approach in order to determine the maternal ECG's estimated value. Findings on synthetic and actual abdominal ECG data demonstrate that the intended technique can extract foetal ECG with signal quality equivalent or superior to that retrieved by multichannel based mechanisms. The anticipated maternal ECG is then nonlinearly matched with the abdominal signal through polynomial networks.