. Author manuscript; available in PMC: 2010 Jul 20.

Published in final edited form as: IEEE Trans Biomed Eng. 2009 Jun 16;57(2):220–237. doi: 10.1109/TBME.2009.2024928

TABLE I.

Evaluation of Inverse Solutions of Annulus Simulation Shown in Fig. 3

	Noise Level	Mesh A	Mesh B	Mesh C

Mean RE	Noise Free	14.80%	10.99%	9.65%
	30dB	15.10%	12.38%	12.20%
	20dB	20.29%	19.75%	19.09%

λ	Noise Free	0.0019	0.0007	0.0003
	30dB	0.0027	0.0015	0.0014
	20dB	0.0144	0.0128	0.0140

MGL Error	Noise Free	1	1	1
	30dB	0.9	0.54	0.42
	20dB	0.70	0.66	0.56

The data represent the arithmetic average of 50 repeated simulations. White noise of 30 and 20 dB SNR is added to the torso-surface measurements before solving the inverse problem. MGL: maximum gradient location in the vector of the inverse solution.