Example script to generate 10 random samples:
python Analytical_data/generate_random_dataset.py -n 10 --type mha --dtype float32 --min_radius 4 --max_radius 32 --min_ratio 1 --max_ratio 60 --min_activity 100 --max_activity 1000 --acquisition_time 15 --radionuclide Lu177 --nspheres 8 --background --sphere 0.7 --ellipse 0.1 --cylinder 0.1 --convex 0.1 --grad_act --nproj 120 --sid 280 --fov 532,388 --attenuationmapfolder attmap_data/ --organlabels rois_labels.json --organratios activity_ratios.json --organproba 0.5 --spect_system siemens-intevo-megp --save_src --rec_fp 10 --project --output_folder output_folder
I know this is not optimal but the --attenuatiomapfolder must contain the following files, for each CT image:
- XXX_attmap_cropped_rot.mhd: the attenution map at the CT resolution in the desired orientation.
- XXX_rois_labels_cropped_rot.mhd: the corresponding ROIs label mask (still at the CT resolution).
- XXX_attmap_cropped_rot_4mm.mhd: the attenuation map with a 4mm resolution, which will be used for reconstructions (rec_fp)
The --organlabels json file specifies the labels contained in the ROIs label masks (XXX_rois_labels_cropped_rot.mhd images). It should be of the following form:
{
"body": 1,
"liver": 2,
"kidney_left": 3,
"kidney_right": 4,
"spleen": 5,
"gallbladder": 6,
"stomach": 7,
"pancreas": 8,
"small_bowel": 9,
"colon": 10,
"duodenum": 11,
"urinary_bladder": 12
}
The --organratios json file should contain [minTBR, maxTBR] values, in the following form:
{
"body": [1,1],
"liver": [6,20],
"kidney_left": [20,60],
"kidney_right": [20,60],
"spleen": [20,60],
"gallbladder": [2,8],
"stomach": [2,8],
"pancreas": [2,8],
"small_bowel": [2,8],
"colon": [2,8],
"duodenum": [2,8],
"urinary_bladder": [6,20]
}
python Simu_data/opengate/gaga_garf_optimize.py -a 5e8 --like_img data/source_4mm_MBq.mha --projections data/FantomeSpheresSansBDF_PW_minus_y.mha --ct data/ct_4mm.mha --radionuclide Lu177 --batchsize 2e5 --gan_pth output_gan_training_1e8/gan_1e8_200epochs_1e5b.pth --garf_pth output_garf_training_2e9_rr5000/garf_lu177_melp_rr5000_JZ_2000_5L_512H.pth --device gpu --geom data/geom_dicom.xml --nprojs 120 --nsubsets 8 --output_folder output_rec_gaga_garf/ --axis y --acquisition_time 15 --with_gan