The commonly used model of the heart for medical applications suffers from some incompleteness when explaining different kinds of measured forces in vivo studies by medical experts. In this paper, we make a statistical analysis of the so-called angle of intrusion automatically. The basis of the proposed method is a set of histological preparations showing heart fibre tissue. We adapt a multi-scale midline extraction process to extract the myocyte strings out of these images and measure the angles of intrusion. Furthermore, a statistical model is derived and validated by the result of a novel parameter estimation technique.