Eigenvector expansion techniques for the inverse problem of electrocardiography applied to a realistic heart-torso geometry

In this paper, we compare various eigenvector expansion techniques for a realistic heart-torso geometry using experimental data studied by other researchers in inverse electrocardiography. With this model, the recently proposed Generalized Eigensystem (GES) approach and the modified GES (tGES) technique perform better than singular value decomposition (SVD) and a modified SVD (tSVD) technique. The tGES and tSVD techniques are combination methods which constrain the solutions like zero-order Tikhonov regulanzation but allow the truncation of modes like GES and SVD.