A Cramér Rao Bound Approach for Evaluating the Quality of Experimental Setups in Electrical Impedance Tomography

Parham Hashemzadeh, Panagiotis Kantaritzis, Panos Liatsis, Richard Bayford, and Sven Nordebo


Electrical Impedance Tomography, Cramer Rao Bound, Fisher Information Matrix, Biomedical Imaging


In this report, we propose the application of the Cramer Rao Lower Bound (CRLB) as a performance measure for optimal design of experimental setups in electrical impedance tomography. In particular, we focus on the optimum positioning of electrodes. Cramer Rao Bound is bounded from below by the inverse of the Fisher information matrix (FIM). FIM incorporates all aspects of the forward problem, statistical properties of the measurement noise, and multi-frequency data. We consider the application of CRB in both the deterministic as well as the Bayesian setting. We first present the CRB for the case of the unbiased estimator and then the Bayesian Cramer Rao Bound (BCRB) for the case of the biased estimator. All CRB computations are performed using a measured noise model from a clinical experiment.

Important Links:

Go Back