Efficient sparse image reconstruction using l1/ 2-regularization for cardiac MRI

摘要

Compressed Sensing (CS) [1] is used in MRI for accelerating the imaging speed by highly under-sampling the acquired data. Sparsity in CS is used for finding a unique solution to the ill-conditioned problem of equations, caused by the compressive under-sampling of data. In the recent past, various approaches have been proposed for sparsity minimization using different lp(0 ≤ p ≤ 1) norms. In this paper, l_(1/2)-regularization based CS reconstruction algorithm for MR images is proposed.