International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056 Volume: 05 Issue: 07 | July 2018 www.irjet.net p-ISSN: 2395-0072 © 2018, IRJET | Impact Factor value: 7.211 | ISO 9001:2008 Certified Journal | Page 2577 Fetal ECG Extraction using Wavelet Transform Neha Para 1 , Dr. S. Wadhawani 2 1 M.E., Dept. of Electrical Engineering, MITS Gwalior, M.P., India 2 Professor, Dept. of Electrical Engineering, MITS Gwalior, M.P., India ---------------------------------------------------------------------***--------------------------------------------------------------------- Abstract - Monitoring of fetal is the way to find the condition of unborn baby in the womb during pregnancy, labor and delivery period by continuous monitoring of fetal heart rate. Measurement of electrical activity of heart is called ECG i.e. Electrocardiography. This monitoring measures and records the electrical activity of the heart of the fetal and provide important details about fetal heart. Normal heart rate of the fetal shows the safe birth of the baby. The thorax (chest) ECG contains only maternal ECG (MECG) but abdominal ECG is a composite ECG signal containing both mother as well as fetal ECG. Since past, in this field lots of research work have been performed which uses threshold and filtering method along with neural network. This paper presents an algorithm which uses Wavelet Transform to extract fetal heart rate from an abdominal ECG signals which is measured at mother’s abdomen. Wavelet Transform is very effective and most popular method for detecting the characteristics of non stationary signals. The proposed algorithm is implemented by in three steps: 1) The abdominal ECG signal (AECG) is recorded from mother’s abdomen and decomposed to estimate maternal ECG signal using wavelet Transform. 2) ECG of fetal (FECG) is extracted by subtracting MECG from AECG signal. 3) Then from extracted FECG signal, R-peaks are identified to obtain fetal heart rate. This research is implemented on 15 recorded signals taken from Physionet using MATLAB which shows consistency of the proposed approach. Key Words: ECG(Electrocardiography), AECG(Abdominal Electrocardiogram),MECG(Mother Electrocardiography), FECG(Fetal Electrocardiography), FHR(Fetal Heart Rate), MHR(Maternal Heart Rate), Wavelet Transform, Mean RR interval, Non- Invasive technique. 1. INTRODUCTION Monitoring of fetal heart is the technique to find the health- status and possible heart diseases during pregnancy. Fetal heart rate (FRH) is the one of the most important tool which is used nowadays in clinical investigations to examine the heart status or health state of fetus during pregnancy and labor. FRH is the mostly used parameter during pregnancy. Monitoring of fetal heart rate helps to detect changes in normal heart rate and yields information about the condition of baby during pregnancy or labor. If any changes are detected, doctors can take some steps to treat the underlying problems. Fetal ECG (FECG) are generally much weaker than maternal ECG (MECG) but fetal heart rate (FHR) is higher than maternal heart rate (MHR) [1]. Eight out of one thousand born live infants have some form of heart defect, making it the single most common class of congenital abnormalities. Identification of the heart defects during early pregnancy reduces risks by timely treatment or planned delivery. Some defects are major while others are minor. They can effect baby’s physical and mental development and appearances. During first trimester of pregnancy most of the heart defects are detected. Detection of heart rate of the fetal can be useful for other cases like: detection of gestational age, monitor the progress of labor, miscarriage rate detection and evaluation and monitoring of uterine contraction etc. To monitor the fetal during pregnancy techniques are classified into invasive and non- invasive methods. Invasive methods generally more risky to the fetal because this method involves needles and probes inserted into uterus. But it can be done from about 14 weeks to 20 weeks gestation. In the non-invasive Fetal ECG (FECG) monitoring, surface electrodes which is placed on abdominal surface of mothers provides valuable information about the cardiac electrical activity of fetus in the womb. By using traditional techniques, extraction of the FECG from the abdominal signal is very hard even not possible. Electrocardiogram (ECG) is the simplest and painless technique to find heart rate. It is non-invasive method to estimate fetal heart condition. ECG provides heart (cardiac) waveform. This cardiac waveform of fetal helps to the physicians to find out or diagnose the arrhythmia such as Bradycardia, Tachycardia, Congenital heart diseases, Asphyxia and Hypoxia. In non-invasive technique the electrodes are placed on the abdominal surface of the pregnant lady or mother, then AECG is recorded which is the combination of the mother and fetal ECG with some noises. It takes great amount of noises called as artifacts. In the ECG signal measuring process four main types of noise and interference occur such as 50 Hz power line interference, Base line drift, Maternal electromyogram (EMG) noise and random Electronic noise [2]. The original FECG signal is very complex and severely contaminated by external disturbances or noises. The object of our work is to remove the noises in signals which are occurred due to power line interference, movement of patient etc. Numerous methods and different approaches have been proposed and used for fetal electrocardiogram (FECG) signal detection: Wavelet based ECG extraction methods [1][2][3][4][5], An Automated Methodology for Fetal Heart Rate Extraction from the Abdominal Electrocardiogram [8], least mean squares adaptive noise cancellation filtering [10],