Denoising ECG signals

Abstract

Extraction of fetal Electrocardiogram (ECG) signal has always been a problem, due to the fetal heart being small, the electric current it generates is rather low causing it to have a low signal-noise ratio as compared to the maternal ECG (mECG). ECG signal is known to have plenty of noise signal that corrupts its data, some of these noises are baseline wandering (BW), power-line interferences and electromyographic interferences (EMG). These noises present in the ECG signal make it difficult for clinicians to perform a proper diagnosis to identify fetal abnormalities. To tackle this problem, we have implemented the Discrete Wavelet Transform (DWT) method to remove the above noise from a noisy maternal ECG signal. The ECG signals were obtained from the fetal ECG synthetic Database while the other noise signals were obtained from the MIT-BIH Noise Stress Test Database.