Joint energy and Laplacian regularization in the inverse problem of electrocardiography

Abstract

In the inverse problem of electrocardiography one attempts to characterize cardiac electrical activity based on noninvasive measurements on the torso surface and knowledge of body geometry. This problem is ill-conditioned and requires regularization. Traditional methods minimize a spatial error at each time instant or a temporal error at each spatial location. Here the authors describe a method which attempts to use more than one spatial constraint in the regularization of the inverse problem at each time instant.<>