def CT(load_path, dset, sc, ang_freq, Exp_bin, bin_param, nTrain, nTD, nVal,
nVD, bpath):
dataset = ddf.load_and_preprocess_real_data(load_path, dset, sc)
meta = ddf.load_meta(load_path + dset + '/', sc)
pix_size = meta['pix_size']
src_rad = meta['s2o']
det_rad = meta['o2d']
data_obj = ddf.real_data(dataset,
pix_size,
src_rad,
det_rad,
ang_freq,
zoom=False)
CT_obj = ddf.CCB_CT(data_obj)
CT_obj.init_algo()
spf_space, Exp_op = ddf.support_functions.ExpOp_builder(
bin_param, CT_obj.filter_space, interp=Exp_bin)
# Create the FDK binned operator
# CT_obj.FDK_bin_nn = CT_obj.FDK_op * Exp_op
# Create the NN-FDK object
CT_obj.NNFDK = nn.NNFDK_class(CT_obj,
nTrain,
nTD,
nVal,
nVD,
Exp_bin,
Exp_op,
bin_param,
dset=dset,
base_path=bpath)
CT_obj.rec_methods += [CT_obj.NNFDK]
return CT_obj
def load_and_preprocess(path, dset, redo, ang_freq):
dataset, vecs = cap.load_and_preprocess(path,
dset,
redo=redo,
ang_freq=ang_freq)
meta = ddf.load_meta(path + dset + '/', 1)
return dataset, vecs, meta
def CT(load_path, dset, sc, ang_freq):
dataset = ddf.load_and_preprocess_real_data(load_path, dset, sc)
meta = ddf.load_meta(load_path + dset + '/', sc)
pix_size = meta['pix_size']
src_rad = meta['s2o']
det_rad = meta['o2d']
data_obj = ddf.real_data(dataset, pix_size, src_rad, det_rad, ang_freq,
zoom=False)
CT_obj = ddf.CCB_CT(data_obj)
CT_obj.init_algo()
return CT_obj
def CT(path, dset, sc, ang_freq, vox):
dataset, vecs = cap.load_and_preprocess(path, dset, ang_freq=ang_freq)
meta = ddf.load_meta(f'{path}{dset}/', 1)
pix_size = sc * meta['pix_size']
src_rad = meta['s2o']
det_rad = meta['o2d']
data_obj = ddf.real_data(dataset,
pix_size,
src_rad,
det_rad,
ang_freq,
vox=vox,
vecs=vecs)
CT_obj = ddf.CCB_CT(data_obj)
CT_obj.init_algo()
return CT_obj
def CT(load_path, dset, sc, ang_freq, Exp_bin, bin_param, offset):
dataset = ddf.load_and_preprocess_real_data(load_path, dset, sc)
meta = ddf.load_meta(load_path + dset + '/', sc)
pix_size = meta['pix_size']
src_rad = meta['s2o']
det_rad = meta['o2d']
data_obj = ddf.real_data(dataset,
pix_size,
src_rad,
det_rad,
ang_freq,
zoom=True,
offset=offset)
CT_obj = ddf.CCB_CT(data_obj)
CT_obj.init_algo()
CT_obj.init_DDF_FDK()
return CT_obj
def CT(path, dset, sc, ang_freq, Exp_bin, bin_param, nTrain, nTD, nVal, nVD,
vox, bp):
dataset, vecs = cap.load_and_preprocess(path, dset, ang_freq=ang_freq)
meta = ddf.load_meta(f'{path}{dset}/', 1)
pix_size = sc * meta['pix_size']
src_rad = meta['s2o']
det_rad = meta['o2d']
data_obj = ddf.real_data(dataset,
pix_size,
src_rad,
det_rad,
ang_freq,
vox=vox,
vecs=vecs)
CT_obj = ddf.CCB_CT(data_obj)
CT_obj.init_algo()
spf_space, Exp_op = ddf.support_functions.ExpOp_builder(
bin_param, CT_obj.filter_space, interp=Exp_bin)
# Create the FDK binned operator
# CT_obj.FDK_bin_nn = CT_obj.FDK_op * Exp_op
# Create the NN-FDK object
CT_obj.NNFDK = nn.NNFDK_class(CT_obj,
nTrain,
nTD,
nVal,
nVD,
Exp_bin,
Exp_op,
bin_param,
dset=dset,
base_path=bp)
CT_obj.rec_methods += [CT_obj.NNFDK]
return CT_obj
filt = filt * (np.cos(freq * np.pi / 2)**2) / 2 / w_detu
# filt = filt / 2 / w_detu
return filt
# %%
path = '/bigstore/lagerwer/data/FleXray/pomegranate1_02MAR/'
#path = '/export/scratch2/lagerwer/data/FleXray/walnuts_10MAY/walnut_01/'
dset = 'noisy'
#dset2 = 'good'
pd = 'processed_data/'
sc = 1
ang_freq = 1
redo = False
dataset = ddf.load_and_preprocess_real_data(path, dset, sc, redo=redo)
meta = ddf.load_meta(path + dset + '/', sc)
pix_size = meta['pix_size']
src_rad = meta['s2o']
det_rad = meta['o2d']
#dataset = {'g': }
# Specifics for the expansion operator
Exp_bin = 'linear'
bin_param = 2
# %% Create a test phantom
# Create a data object
t2 = time.time()
offset = 1.26 / 360 * 2 * np.pi
data_obj = ddf.real_data(dataset,
def load_and_preprocess(path, dset, sc, redo, ang_freq):
dataset, vecs = cap.load_and_preprocess(path, dset, ang_freq=ang_freq)
meta = ddf.load_meta(f'{path}{dset}/', 1)
return dataset, meta, vecs
def compute_GS_and_mask(path, redo):
save_path = f'{path}processed_data/'
niter = 50
sc = 2
vecs = np.zeros((0, 12))
g = np.zeros((0, 768, 972))
for i in [1, 2, 3]:
dset = f'tubeV{i}'
lp = f'{path}{dset}/'
lpvecs = f'{lp}scan_geom_corrected.geom'
vecs = np.append(vecs, np.loadtxt(lpvecs)[:-1, :] / 10, axis=0)
g = np.append(g,
preprocess_data_portrait(path, dset, redo=redo),
axis=0)
if i == 2:
meta = ddf.load_meta(lp, 1)
rec = do_AGD(meta, vecs, sc, g, niter)
vox = np.shape(rec)[0]
save = np.zeros((3, vox, vox))
save[0, :, :], save[1, :, :] = rec[:, :, vox // 2], rec[:, vox // 2, :]
save[2, :, :] = rec[vox // 2, :, :]
np.save(f'{save_path}rec_ax_AGD50', save)
# %%
np.save(f'{save_path}rec_AGD50', rec)
end = time.time()
# %%
rec *= (rec > 0.03)
edge = vox // 32
edge_t = np.zeros(np.shape(rec))
edge_t[edge:-edge, edge:-edge, edge:-edge] = 1
rec *= edge_t
del edge_t
save[0, :, :], save[1, :, :] = rec[:, :, vox // 2], rec[:, vox // 2, :]
save[2, :, :] = rec[vox // 2, :, :]
np.save(f'{save_path}GT_ax', save)
diter = int(1.5 * 2**(np.log2(vox) - 5))
it = 5
mask = sp.binary_erosion(rec, iterations=it)
mask = sp.binary_dilation(mask, iterations=diter + it)
save[0, :, :], save[1, :, :] = mask[:, :, vox // 2], mask[:, vox // 2, :]
save[2, :, :] = mask[vox // 2, :, :]
np.save(f'{save_path}mask_ax', save)
# %%
np.save(f'{save_path}ground_truth.npy', rec)
np.save(f'{save_path}mask.npy', mask)
t3 = time.time()
print(t3 - end, 'Finished computing mask and ground truth')
# %%
#path = '/export/scratch2/lagerwer/data/FleXray/Walnuts/Walnut1/Projections/'
#dset = f'tubeV{2}'
#pd = 'processed_data/'
#ang_freq = 1
#redo = False
#sc = 4
#
#load_and_preprocess(path, dset, redo=redo)
def load_and_preprocess(path, dset, sc, redo):
dataset = ddf.load_and_preprocess_real_data(path, dset, sc, redo=redo)
meta = ddf.load_meta(path + dset + '/', sc)
return dataset, meta