Search

Impedance cardiography (ICG) is an attractive noninvasive method for measuring stroke volume and cardiac output. However, it is easily disturbed by the artifacts such as respiration and body shaking. Considering the correlation between physiological and motion signals, and the synchronization relationship among the physiological signals, this paper introduces an ICG denoising method based on canonical correlation analysis (CCA) and coherence analysis to remove the artifacts during the measurement. First, the CCA extracts the shared components between the electrocardiogram (ECG) and the motion signal, as well as the shared components between ICG and the motion signal. Then, set those shared components to zero to suppress the primary motion artifacts. Next, the coherence analysis was used to calculate the synchronization relationship between obtained ECG and ICG components. Finally, the components with strong synchronization relationships were used to reconstruct the ICG signal. The denoising method was evaluated for 54 subjects during lying and walking. Experimental results show that after removing the artifacts, the signal quality index beat contribution factor (BCF) was increased from the original 78.1% to 97.8%, and the physiological parameters measured based on the proposed method were in good agreement with those measured by the standard instrument. The proposed denoising method effectively improves the reliability of analysis and diagnosis on cardiovascular diseases relying on ICG signals. [ABSTRACT FROM AUTHOR]

Impedance cardiography (ICG) is an attractive noninvasive method for measuring stroke volume and cardiac output. However, it is easily disturbed by the artifacts such as respiration and body shaking. Considering the correlation between physiological and motion signals, and the synchronization relationship among the physiological signals, this paper introduces an ICG denoising method based on canonical correlation analysis (CCA) and coherence analysis to remove the artifacts during the measurement. First, the CCA extracts the shared components between the electrocardiogram (ECG) and the motion signal, as well as the shared components between ICG and the motion signal. Then, set those shared components to zero to suppress the primary motion artifacts. Next, the coherence analysis was used to calculate the synchronization relationship between obtained ECG and ICG components. Finally, the components with strong synchronization relationships were used to reconstruct the ICG signal. The denoising method was evaluated for 54 subjects during lying and walking. Experimental results show that after removing the artifacts, the signal quality index beat contribution factor (BCF) was increased from the original 78.1% to 97.8%, and the physiological parameters measured based on the proposed method were in good agreement with those measured by the standard instrument. The proposed denoising method effectively improves the reliability of analysis and diagnosis on cardiovascular diseases relying on ICG signals. [ABSTRACT FROM AUTHOR]