Spike artifact reduction in nonconvex Compressed Sensing

Compressed sensing (CS), a reconstruction method for undersampled MR data, was recently introduced [1]. Since only undersampled data are acquired, CS allows a significant reduction in the time needed for MR experiments. The basic requirement for CS, however, is sparsity in the data. The lack of F background signal in living tissue leads to an intrinsically sparse signal distribution in the F image domain. This makes F MR a suitable application for CS [2]. However, F MR data often suffers from a low SNR, which is problematic for CS. Thus, spike artifacts often appear highly pronounced especially in nonconvex CS reconstructions of noisy F MR data. The present study focuses on the reduction of spike artifacts in these CS reconstructions. Therefore, a post-processing "de-spike algorithm" is proposed, using the fact that the spatial position of spike artifacts depends on the chosen sampling pattern. Numerical phantom simulations as well as exand in-vivo F CSI experiments were performed.