Apply ICA (Independent Component Analysis) to Removing Respiratory Interference from ECG*
*This work was supported in part by The Province-Ministry Joint Key Laboratory of Electromagnetic Field and Electrical Apparatus Reliability, Hebei University of Technology, Tianjin, China.
Removing the respiratory signal is a crucial topic to the high-quality ECG. But, not all models are available in the project. The use of digital filtering, signal averaging, adaptive processing and wavelet transform to remove the respiratory interference have some problems. The ICA algorithm for cancellation of respiratory interference is proposed. It is found that this method is more available to reconstruct high-quality ECG and de-noising artifacts comparing with the wavelet transform. Three steps are performed in the paper. From the simulation aspect, and from the evolution for the ability of de-noising the respiratory signal, as well as from reconstructing ECG, the comparisons between the results using ICA and that using wavelet transform are fulfilled. It is shown that the ICA algorithm is more powerful and more effective to de-noising the respiratory signal from ECG, almost not destroying the original ECG.
