LAUSR

Abstract Markov chain Monte Carlo algorithms are used to sample complex distributions and, as such, are suitable for Baesyan reconstructions of inverse problems. Electrical impedance tomography reconstructions pose such problems, but the evaluation of their forward models are often computationally too expensive for Markov chain Monte Carlo sampling. Herein is proposed a new Markov chain Monte Carlo sampling algorithm based on incomplete evaluation of electrical impedance tomography forward models. Such evaluation greatly reduces the sampling process computational cost, while still providing representative electrical impedance tomography reconstructions.