A fast regularized CT reconstruction method based on a preconditioning matrix for limited-angle fan-beam CT

Abstract

This paper proposes a fast regularized method for limited angle CT, the preconditioning matrix, which includes a weighting matrix and a filtering matrix, is utilized to improve the convergence speed. The projection and back-projection in the iteration process are replaced by the analytic operators to accelerate the computation speed. In addition, a Hamming window function is added to further suppress the noise and artifacts. Real data experiments show that the proposed method outperforms the other three methods in preserving the edge, convergence speed and computing speed.