Bayesian Reconstruction Algorithm for Low-dose CT Based on New Nonlocal Prior Model

In order to improve the quality of low-dose CT reconstructed image, this study proposes a projection symmetry-based modified nonlocal prior model based on the traditional nonlocal prior model. Then, a Bayesian reconstruction algorithm is built combined with this prior model, and it is applied to the noise removal of the low-dose CT projection data. The reconstructed images are obtained by the filtered back-projection(FBP)algorithm. The results of simulated experiment show the proposed algorithm, compared with the algorithms based on the traditional priors, can achieve a superior balance between suppressing noise and preserving edges.