The muscles of the heart are associated with the flow of currents on tissue in the volume conductor formed inside the human body. The electrical currents of the human heart generate a tiny magnetic field beyond the thorax surface. The functional waves generated in terms of the field are detected using Magnetocardiogram (MCG) at the detector level. One of the challenging tasks in the research area is to compute algorithms which help in visualizing/localizing the cardiac anomalies non-invasively at the heart level. In order to localize, a generic model based on prior assumptions is designed which is known as a forward problem. This model depicts the spatial relation between the heart locations and the detectors.

In this paper, a novel method is proposed in the construction of the forward matrix; where the prior heart vectors are constrained with vectorcardiography (VCG) signals. The spatial matrix is subsequently used to estimate the desired position on the myocardium in an inverse way. To evaluate the model, several true prior dipoles are placed on desired positions of the heart each at a time and the inverse problem is solved using multi-start downhill simplex search. The spatial matrices are updated based on VCG during each movement of the test dipoles' positions in the inverse search. The inverse studies are computed and compared between the proposed and the existed methods for the entire cardiac cycle. The performance of the inverse model with unit VCG orientations exhibited an improvement of   mm average localization error than the conventional one.