Classification of cardiac diseases based on singular value decomposition of the magnetocardiogram

For the assessment of cardiac risk factors and the prediction of cardiac risk an automatic classification of different cardiac diseases is necessary. We applied singular value decomposition and subsequent template and power computations to high resolution MCG (magnetocardiogram) mapping of 21 subjects. The results demonstrate the possibility to establish automatic classification for cardiac diseases based on MCG recordings. Larger study groups are necessary for statistical evidence. The results of this study will be implemented into a MCG based cardiac screening software.