Compressive imaging using approximate message passing and a Markov-tree prior

摘要: We propose a novel algorithm for compressive imaging that exploits both the sparsity and persistence across scales found in 2D wavelet transform coefficients of natural images. Like other recent works, we model structure using hidden Markov tree (HMT) but, unlike ours is based on loopy belief propagation (LBP). For LBP, adopt recently proposed “turbo” message passing schedule alternates between exploitation HMT compressive-measurement structure. latter, leverage Donoho, Maleki, Montanari's approximate (AMP) algorithm. Experiments with large image database suggest that, relative to existing schemes, our turbo LBP approach yields state-of-the-art reconstruction performance substantial reduction complexity.