Deep neural networks for inverse problems with pseudodifferential operators: an application to limited-angle tomography

Tatiana A. Bubba, Mathilde Galinier, Matti Lassas +3

We propose a novel convolutional neural network (CNN), called $\Psi$DONet, designed for learning pseudodifferential operators ($\Psi$DOs) in the context of linear inverse problems. Our s…