Fig. 9.

Dataset A (2D): comparing our approach to competitive methods: i/sMAPEM, a regularized tomographic reconstruction method designed to achieve isotropic resolution; ii/ the Moisan’s method designed to extrapolate missing regions in Fourier space; iii/BFLY, a filter designed to reduce MW artifacts. The sMAPEM method takes projections as input, therefore we used the same projections to produce the 2D input (via WBP) for the other methods. On the bottom we display the PSNR and CCC scores obtained for all tested methods.