ECG Baseline Wander Reduction Using Empirical Mode Decomposition: A Case Study on Preterm Infants With Bradycardia

Reducing the baseline wander (BW) distortion in electrocardiogram (ECG) signals while preserving the principal information about the heart rate (HR) is crucial. This paper proposes a method based on empirical mode decomposition (EMD) to mitigate BW distortion in ECG signals. The presented method reduces the BW by removing the residue as well as two low-frequency intrinsic mode functions (IMFs) from the original raw ECG. To investigate the usefulness of the proposed EMD-based BW reduction method, we conducted MATLAB simulations on bradycardia ECG data from the publicly available Preterm Infant Cardio-respiratory Signals (PICS) database. The results demonstrated the effectiveness of the proposed method compared to the traditional Butterworth high-pass filter (HPF). In particular, the proposed method provided a signal-to-noise ratio (SNR) that is 17.58 dB higher than that achieved by the conventional method. This commendable performance confirms the potential of the presented BW denoising approach, which aids in achieving accurate ECG-based preterm neonatal HR detection.