Feature Reconstruction from Incomplete Tomographic Data without Detour

In this paper, we consider the problem of feature reconstruction from incomplete X-ray CT data. Such data problems occur when number measured X-rays is restricted either due to limit radiation exposure or practical constraints, making detection certain rays challenging. Since image a severely ill-posed (unstable) problem, reconstructed images may suffer characteristic artefacts missing features, thus significantly complicating subsequent processing tasks (e.g., edge segmentation). introduce framework for robust convolutional features directly without need computing first. Within our framework, use non-linear variational regularization methods that can be adapted variety and several limited situations. The proposed method minimizes an energy functional being sum dependent data-fitting term additional penalty accounting specific properties features. numerical experiments, instances reconstructions angular under-sampled show approach able reliably reconstruct maps in case.