Choose, implement, and report a reconstruction method for an ill-posed imaging inverse problem: Tikhonov and total variation, sparsity, plug-and-play and RED, unrolled and learned primal-dual networks, convex and weakly convex learned regularisers, self-supervised and equivariant training, deep image prior, diffusion and Langevin posterior sampling, and nonlinear problems such as impedance and optical tomography.
