Non-Invasive Locating Of Premature Ventricular Contraction Origin With Low Rank/Tv Regularization

Abstract

Most of the calculation methods for the electrocardiograph (ECG) inverse problem are based on priori assumptions of the instantaneous characteristics of the cardiac electrophysiology. In this paper, we have proposed a novel algorithm based on low rank and sparse decomposition (LSD) + total variation (TV) to solve the illposedness of dynamic ECG-inverse problem. The TV constraint filters out the disturbance of the noise and maintains the local smoothness of the potential. The LSD separates the sparse details from the potential background to prevent the potential details from being lost under the effect of smoothing constraint, thereby improving the accuracy of cardiac potential recovery.