Adversarially Learned Iterative Reconstruction for Imaging Inverse Problems

In numerous practical applications, especially in medical image reconstruction, it is often infeasible to obtain a large ensemble of ground-truth/measurement pairs for supervised learning. Therefore, imperative develop unsupervised learning protocols that are competitive with approaches performance. Motivated by the maximum-likelihood principle, we propose an framework solving ill-posed inverse problems. Instead seeking pixel-wise proximity between reconstructed and ground-truth images, proposed approach learns iterative reconstruction network whose output matches distribution. Considering tomographic as application, demonstrate not only performs on par its variant terms objective quality measures, but also successfully circumvents issue over-smoothing tend suffer from. The improvement comes at expense higher training complexity, but, once trained, time remains same counterpart.