A new method for inverse problem of body surface potential mapping-self constraint super resolution (SCSR) method

摘要

The ECG inverse problem is aimed to estimate cardiac electrical activity such as epicardial potentials from measured precordial potential distribution. We investigated the accuracy and stability of ECG inverse solution by applying singular value decomposition and proposed a new method to obtain globally and locally valid regularized solution by unified use of the information criterion (minimum description length principle) and Iakovidis method (1992). As a numerical model for computer validation of this method, numerically-modelled human thorax was used and epicardial (heart surface) potential body surface potential mapping was generated by using boundary element method. We applied our method to this numerical model and found our self constraint super resolution method (SCSR) provides both globally and locally improved solution.