The standard ECG remains the most common noninvasive tool for diagnosing and studying atrial fibrillation (AF). Due to the much higher amplitude of the electrical ventricular activity in the surface ECG, isolation of the atrial activity component is crucial to the analysis and characterization of AF. An average beat subtraction (ABS) based method is developed to perform this QRST cancellation. In contrast with standard methods, two sets of templates are created instead of one: one set for the QRS complexes and one for the T-waves. The QRS complexes are clustered according to their morphology; the T-waves, using both their preceding RR interval and their morphology. Next, spatial optimization (rotation and scaling) is applied to the QRS templates. ECG signals generated by a biophysical model are used to evaluate the performance of the proposed method in comparison with two other QRST cancellation methods. The proposed method decreases the averaged relative error by 19.7% and 29.0% in comparison with the standard ABS and the standard spatiotemporal method, respectively.
展开▼
