Laplacian and Bilateral Weighted Relative Total Variation Sparse Angle CT Reconstruction

Abstract Reconstruction of computed tomography (CT) images under sparse data conditions often leads to compromised quality, which can result in diagnostic inaccuracies. This study introduces a novel iterative reconstruction algorithm that combines a second-order differential Laplacian operator with a bilateral weighted relative total variation model to enhance the CT image quality from sparse datasets. The approach is designed to efficiently capture sharp edges and fine textures while reducing noise and maintaining critical edge features. Numerical simulations and preliminary clinical testing demonstrate that the algorithm significantly reduces streak artifacts and improves edge clarity, outperforming traditional methods in both qualitative and quantitative analyses. In summary, the developed iterative reconstruction algorithm substantially enhances the reconstruction quality of CT images with sparse data, showing significant advantages over conventional techniques, and promises to significantly improve the accuracy and reliability of clinical imaging diagnostics.