Features extraction of atrial activity from surface ECG

In this paper, a novel model of independent component analysis (ICA), based on sparse representation, was proposed to extract atrial activity electrophysiology from 12-leads ECG. Some preprocessing was done upon the acquired 12-leads ECG and the wavelet transform was used to implement the sparse representation of ECG. ICA in wavelet domain was proposed to deal with the transformed coefficients of the 12-leads ECG. The atrial activity electrophysiology, which was separated as an independent source from the transformed coefficients of the 12-leads ECG, was verified by means of spectral analysis. The application of average atrial fibrillation cycle length was discussed in details. Results showed that this new approach had the ability to extract atrial activity electrophysiological signal, having great importance to the mechanisms research of atrial activity electrophysiology.