Learning a microlocal priorfor limited-angle tomography

Digital breast tomosynthesis is an ill posed inverse problem. In this paper, we provide a try to overcome the problem of stretching artefacts of DBT with the help of learning from the microlocal priors.