Some new results on the inverse problem for electrocardiography

The use of several mathematical methods for estimating epicardial electrocardiography (ECG) potentials from arrays of body surface has been reported in the literature; most of these methods are based on the least-squares reconstruction principle and operate in the time-space domain. The authors present some new results of the singular value decomposition (SVD)-based inverse methods using real data and a realistic high-resolution canine anatomy model. A computer-based system was designed to reconstruct an accurate canine geometric model of anatomy from 31 CT-scan cross-sections that include the internal organs and skeletal muscle inhomogeneities.<>