hyper_params={
'filter_type': 'Hann',
'frequency_scaling': 0.8
})
cg_reconstructor = CGReconstructor(ray_trafo, reco_space.zero(), 4)
gn_reconstructor = GaussNewtonReconstructor(ray_trafo, reco_space.zero(), 2)
lw_reconstructor = LandweberReconstructor(ray_trafo, reco_space.zero(), 8)
mlem_reconstructor = MLEMReconstructor(ray_trafo, 0.5 * reco_space.one(), 1)
reconstructors = [
fbp_reconstructor, cg_reconstructor, gn_reconstructor, lw_reconstructor,
mlem_reconstructor
]
options = {'save_iterates': True}
eval_tt.append_all_combinations(reconstructors=reconstructors,
test_data=[test_data],
options=options)
# %% run task table
results = eval_tt.run()
results.apply_measures([PSNR, SSIM])
print(results)
# %% plot reconstructions
fig = results.plot_all_reconstructions(fig_size=(9, 4), vrange='individual')
# %% plot convergence
results.plot_convergence(1, fig_size=(9, 6), gridspec_kw={'hspace': 0.5})
results.plot_performance(PSNR)
hyper_params={
'filter_type': 'Hann',
'frequency_scaling': 0.8
})
fbp_unet_reconstructor = FBPUNetReconstructor(ray_trafo,
batch_size=64,
use_cuda=True)
state_filename = 'fbp_unet_reconstructor_lodopab_baseline_state.pt'
with open(state_filename, 'wb') as file:
r = requests.get('https://github.com/jleuschn/supp.dival/raw/master/'
'examples/'
'fbp_unet_reconstructor_lodopab_baseline_state.pt')
file.write(r.content)
fbp_unet_reconstructor.load_params(state_filename)
reconstructors = [fbp_reconstructor, fbp_unet_reconstructor]
eval_tt.append_all_combinations(reconstructors=reconstructors,
test_data=[test_data],
datasets=[dataset],
options={'skip_training': True})
# %% run task table
results = eval_tt.run()
results.apply_measures([PSNR, SSIM])
print(results)
# %% plot reconstructions
fig = results.plot_all_reconstructions(fig_size=(9, 4), vrange='individual')