Supplementary Material for "Bayesian Experimental Design for Computed Tomography with the Linearised Deep Image Prior"

Supplementary material for the paper "Bayesian Experimental Design for Computed Tomography with the Linearised Deep Image Prior". The source code of the project can be found at github.com/educating-dip/bayesian_experimental_design. Contains experimental results, including tensorboard logs and other saved information: Initial DIP reconstruction and MLL optimisation for linearised DIP prior hyperparameters initial_mll_noise05.zip initial_mll_noise10.zip Selection by linearised DIP mll_prior_ESE_noise05.zip mll_prior_ESE_noise10.zip mll_prior_EIG_noise05.zip mll_prior_EIG_noise10.zip Selection by linearised DIP, with DIP retrained every 5 angles mll_prior_refinement_ESE_noise05.zip mll_prior_refinement_ESE_noise10.zip mll_prior_refinement_EIG_noise05.zip mll_prior_refinement_EIG_noise10.zip Selection by linearised DIP (g-prior) g_prior_ESE_noise05.zip g_prior_ESE_noise10.zip g_prior_EIG_noise05.zip g_prior_EIG_noise10.zip Selection by linearised DIP (g-prior), with DIP retrained every 5 angles g_prior_refinement_ESE_noise05.zip g_prior_refinement_ESE_noise10.zip g_prior_refinement_EIG_noise05.zip g_prior_refinement_EIG_noise10.zip Selection by isotropic model linear_model_isotropic_ESE_noise05.zip linear_model_isotropic_ESE_noise10.zip linear_model_isotropic_EIG_noise05.zip linear_model_isotropic_EIG_noise10.zip linear_model_isotropic_diagonal_EIG_noise05.zip linear_model_isotropic_diagonal_EIG_noise10.zip Selection by Matern-½ process linear_model_gp_ESE_noise05.zip linear_model_gp_ESE_noise10.zip linear_model_gp_EIG_noise05.zip linear_model_gp_EIG_noise10.zip linear_model_gp_diagonal_EIG_noise05.zip linear_model_gp_diagonal_EIG_noise10.zip Baseline reconstruction using equidistant angles equidistant_baseline_noise05.zip equidistant_baseline_noise10.zip Baseline reconstruction using randomly selected angles random_baseline_noise05.zip random_baseline_noise10.zi