ICG signal denoising based on ICEEMDAN and PSO-VMD methods

Area of Science:

• Biomedical Engineering
• Cardiovascular Physiology
• Signal Processing

Background:

• Impedance cardiography (ICG) is vital for assessing cardiac function, but its accuracy relies on precise feature point identification.
• Conventional noise reduction methods can distort critical amplitude and temporal characteristics of ICG signals, impacting parameter calculations.
• Artifacts from noise and breathing can compromise the clinical utility of ICG data.

Purpose of the Study:

• To develop and evaluate a novel noise and artifact elimination method for impedance cardiography (ICG) signals.
• To preserve the amplitude and temporal fidelity of ICG signals for accurate cardiac parameter computation.
• To enhance the reliability of feature point extraction from ICG data.

Main Methods:

• Implementation of Improved Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (ICEEMDAN).
• Application of Particle Swarm Optimization-based Variational Mode Decomposition Algorithm (PSO-VMD) for signal decomposition and noise removal.
• Comparative analysis against wavelet-based methods and Ensemble Empirical Mode Decomposition (EEMD).

Main Results:

• The proposed ICEEMDAN and PSO-VMD method significantly improved the signal-to-noise ratio (SNR).
• The technique demonstrated a lower root-mean-square error (RMSE) compared to existing methods.
• Enhanced correlation and waveform consistency with the original ICG signals were observed.

Conclusions:

• The ICEEMDAN and PSO-VMD approach offers a superior method for processing ICG signals, effectively removing noise and artifacts.
• This technique ensures the preservation of essential signal characteristics, leading to more accurate cardiac function evaluation.
• The findings support the clinical applicability of this advanced signal processing method for ICG analysis.