Including a priori knowledge in the solution of the inverse problem during atrial fibrillation

Regularization methods for the inverse problem face important challenges during atrial fibrillation (AF), so any a priori information available during the electrophysiological study may improve the solution. We propose a Tikhonov-based inverse problem formulation that incorporates extra information provided by noisy intracavitary measurements near the endocardium. Since we introduced two different sources of information (Tikhonov estimation and a priori measurements), two regularization parameters must be tuned using a generalized L-Curve method. Performance of this method is studied using several time-based metrics proposed in previous works. The method proposed improved estimations accuracy in sinus rhythm, simple and complex AF models.