Sinogram restoration for low-dose X-ray computed tomography using regularized Perona–Malik equation with intuitionistic fuzzy entropy

Edges and flat areas in the sinogram of low-dose X-ray computed tomography are difficult to distinguish. This lack of distinction leads to excessive smoothing in the sinogram restoration. To address this problem, we propose a sinogram restoration algorithm using the regularized Perona–Malik (P–M) equation with intuitionistic fuzzy entropy. Firstly, considering the sinogram fuzziness, a novel edge indicator function is constructed using both the gradient magnitude and intuitionistic fuzzy entropy. Secondly, using the constructed edge indicator function as the diffusion coefficient, a novel regularized P–M equation smoothing model is presented. The proposed model overcomes the shortage of traditional P–M equations, which are ill-conditioned. Moreover, it performs diffusion with different directions and intensities in different regions of the sinogram. The optimal solution of the proposed algorithm is obtained by using the additional operator splitting method. Finally, the reconstructed image is achieved by filtered back projection from the smoothed sinogram. Experimental results show that the presented method can retain important edges while smoothing noise and perform better than others.