Wavelet based denoising for suppression of respiratory and motion artifacts in impedance cardiography

Impedance cardiography senses the variation in the thoracic impedance caused by variation in the blood volume and it is used for estimating the stroke volume and other cardiovascular indices. Respiratory and motion artifacts in the sensed signal introduce errors in these estimations. A denoising technique, using discrete Meyer and symlet-26 wavelets, with scale-dependent thresholding for suppressing the respiratory artifact and limiting of the wavelet coefficients for suppressing the motion artifact is investigated. Denoising of signals with simulated respiratory artifacts improved the signal-to-artifact ratio by 23.5 dB. Denoising of signals with real respiratory and motion artifacts resulted in the values of L2 norm and max-min based improvement indices being close to one, indicating effective suppression of artifacts without any significant signal distortion.