Noise and artifact comparison for Fourier and polynomial phase correction used with Fourier reconstruction of asymmetric data sets

Two processes of phase correction, Fourier phase mapping and second-order polynomial phase fitting, are compared in combination with four different schemes for Fourier reconstruction of asymmetric data, using one-dimensional simulations and two-dimensional human head data. Polynomial phase correction provides systematically less image noise and is much less affected by localized phase differences caused by object edges and motion.