Three-dimensional multi-resolution statistical reconstruction for tomosynthesis

In this paper, we propose a novel three-dimensional maximum a posteriori (MAP) algorithm for tomosynthetic reconstruction. The motivation of this work is to improve the image quality while simultaneously reduce the computational cost that is the most challenging problem of the statistical (Bayesian) reconstruction algorithm. The proposed method utilizes a multi-resolution representation for both reconstruction and projections. The imaged object is reconstructed from the coarsest scale to the finest scale, sequentially. Simulations are presented showing that the proposed algorithm produces better imaged-quality and leads to faster convergence than the alternative methods.