Few-view CT imaging is an important approach to reduce the ionizing radiation dose. In this paper, we propose a threedimensional (3D) deep-learning-based method for few-view CT image reconstruction directly from 3D projection data. The large memory requir
展开▼