A hybrid variational mode decomposition framework for enhanced cardiac output estimation using impedance cardiography

Area of Science:

• Biomedical Engineering
• Cardiovascular Physiology
• Signal Processing

Background:

• Accurate cardiac output (CO) estimation is critical for managing cardiovascular disorders.
• Impedance cardiography (ICG) signals are prone to noise artifacts, limiting reliability.
• Existing denoising methods struggle with complex noise in ICG data.

Purpose of the Study:

• To develop and validate a novel three-stage denoising framework for enhancing ICG signal quality.
• To improve the accuracy and robustness of cardiac output estimation from denoised ICG signals.
• To assess the framework's performance against state-of-the-art methods and its clinical feasibility.

Main Methods:

• A three-stage denoising framework integrating Variational Mode Decomposition (VMD), Non-Local Means (NLM), and Discrete Wavelet Transform (DWT).
• Validation on the ReBeatICG dataset, including signals with motion artifacts and baseline drift.
• Performance evaluation using metrics like Signal-to-Noise Ratio (SNR), Mean Squared Error (MSE), Percent Root Mean Square Difference (PRD), F1-score, and Denoising Robustness Index (DRI).

Main Results:

• Achieved up to 1.2 dB SNR improvement and reduced MSE by 13% and PRD by 9% compared to two-stage methods.
• Enhanced fiducial point detection (up to 4.4% F1-score increase) and preserved heart rate variability (HRV) fidelity (0.91 correlation coefficient).
• Demonstrated superior denoising robustness and signal fidelity preservation under various noise conditions, with statistical validation.

Conclusions:

• The proposed VMD-NLM-DWT framework significantly enhances ICG signal quality for robust cardiac output estimation.
• The method outperforms existing techniques in preserving clinically relevant signal features and accuracy.
• Computational efficiency supports real-time application in clinical and ambulatory cardiovascular monitoring.