Curvature weighted total variation minimization for limited angle CT reconstruction

Regularization methods based on image prior information are the mainstream approach for addressing limited-angle problems. In limited-angle CT reconstruction, the missing data in continuous angle range scanning can disrupt the continuity of surface curvature in the image, which is one of the significant impacts. In this paper, we propose a limited-angle CT reconstruction algorithm based on weighted total variation (TV) minimization using surface curvature. This algorithm assigns different weights to different directions of the image based on surface curvature information, thereby utilizing the anisotropic properties. Subsequently, we introduce the alternating minimization technique to solve the proposed model. Experimental results demonstrate that our proposed method is highly efficient and outperforms other competing algorithms used in limited-angle CT reconstruction in terms of noise suppression and edge preservation.