Two mechanisms for electrocardiographic deconvolution

Electrocardiographic imaging requires computation of a time-varying “source” (e.g. epicardial potentials) from measurements (body surface potentials) with a known linear dependence on the source. The problem has been usually treated in a time-independent fashion, but since it is ill-posed, and since solutions at adjacent time points must be codependent, it has been suggested that time-dependent information be incorporated into the regularization procedure. The authors supply two different mechanisms for incorporating time in a natural fashion based on treatment of the data as a generalized convolution