Computational imaging benefits from untrained neural network

In a recent study, investigators from the Chinese Academy of Sciences described how they combined an untrained neural network and physics knowledge to eliminate the limitations of deep-learning-based CI methods.