Comparison of 2D and 3D total variation minimization methods in breast tomosynthesis imaging

In tomosynthesis imaging, a small number of projections are acquired from a limited scan angle which is insufficient to reconstruct the image without undesired artifacts. Iterative reconstruction algorithms have been widely used in order to combat this problem. In this study, an effective compressed sensing (CS) based iterative reconstruction algorithm was implemented by applying total variation minimization in TV_2D and TV_3D forms. ART + TV_2D has shown superior results over ART. However, the effect of regularization in the axial dimension is missing in TV_2D. A 3D phantom which roughly simulates a breast tomosynthesis image was designed to evaluate if TV_3D has a superiority over ART and ART + TV_2D in the sense of root mean square error (RMSE) of a specific layer of interest (LOI) and the entire phantom. Computer simulations show that ART + TV_3D method substantially enhances the reconstructed image by generating fewer artifacts with smaller error rates.