CTprintNet: An Accurate and Stable Deep Unfolding Approach for Few-View CT Reconstruction

In this paper, we propose a new deep learning approach based on unfolded neural networks for the reconstruction of X-ray computed tomography images from few views. We start model-based in compressed sensing framework, described by minimization least squares function plus an edge-preserving prior solution. particular, proposed network automatically estimates internal parameters proximal interior point method solution optimization problem. The numerical tests performed both synthetic and real dataset show effectiveness framework terms accuracy robustness with respect to noise input sinogram when compared other different data-driven approaches.