def __init__(self, geometry):
"""
Parameters
----------
geometry : GeometryCone3D
The geometry used for reconstruction
"""
self.geometry = geometry
self.cosine_weight = tf.get_variable(
name='cosine_weight',
dtype=tf.float32,
initializer=ct_weights.cosine_weights_3d(geometry),
trainable=True)
self.recon_filter = tf.get_variable(
name='recon_filter',
dtype=tf.float32,
initializer=filters.ram_lak_3D(geometry),
trainable=True)
self.relu_alpha = tf.get_variable(
name='relu_alpha',
shape=(1),
dtype=tf.float32,
initializer=tf.constant_initializer(0),
trainable=True)
def __init__(self, geometry):
"""
Parameters
----------
geometry : GeometryCone3D
The geometry used for reconstruction
"""
self.geometry = geometry
self.initializer = tf.contrib.layers.xavier_initializer()
self.cosine_weight = tf.get_variable(name='cosine_weight', dtype=tf.float32,
initializer=ct_weights.cosine_weights_3d(geometry),
trainable=True)
self.recon_filter = tf.get_variable(name='recon_filter', dtype=tf.float32,
initializer=filters.ram_lak_3D(geometry),
trainable=True)
self.relu_alpha = tf.get_variable(name='relu_alpha', shape=(1), dtype=tf.float32,
initializer=tf.constant_initializer(0),
trainable=True)
# for multi perceptron if needed
self.mlp_all = tf.get_variable(name='mlp_all', dtype=tf.float32, initializer=tf.ones(1), trainable=False)
self.mlp_one = tf.get_variable(name='mlp_one', dtype=tf.float32, initializer=tf.ones(1), trainable=False)
self.mlp_two = tf.get_variable(name='mlp_two', dtype=tf.float32, initializer=tf.ones(1), trainable=False)
def example_cone_3d():
# ------------------ Declare Parameters ------------------
# Volume Parameters:
volume_size = 256
volume_shape = [volume_size, volume_size, volume_size]
volume_spacing = [0.5, 0.5, 0.5]
# Detector Parameters:
detector_shape = [2*volume_size, 2*volume_size]
detector_spacing = [1, 1]
# Trajectory Parameters:
number_of_projections = 360
angular_range = 2 * np.pi
source_detector_distance = 1200
source_isocenter_distance = 750
# create Geometry class
geometry = GeometryCone3D(volume_shape, volume_spacing, detector_shape, detector_spacing, number_of_projections, angular_range, source_detector_distance, source_isocenter_distance)
geometry.set_trajectory(circular_trajectory.circular_trajectory_3d(geometry))
# Get Phantom 3d
phantom = shepp_logan.shepp_logan_3d(volume_shape)
# Add required batch dimension
phantom = np.expand_dims(phantom, axis=0)
# ------------------ Call Layers ------------------
# The following code is the new TF2.0 experimental way to tell
# Tensorflow only to allocate GPU memory needed rather then allocate every GPU memory available.
# This is important for the use of the hardware interpolation projector, otherwise there might be not enough memory left
# to allocate the texture memory on the GPU
gpus = tf.config.experimental.list_physical_devices('GPU')
if gpus:
try:
for gpu in gpus:
tf.config.experimental.set_memory_growth(gpu, True)
except RunetimeError as e:
print(e)
sinogram = cone_projection3d(phantom, geometry)
reco_filter = ram_lak_3D(geometry)
sino_freq = tf.signal.fft(tf.cast(sinogram,dtype=tf.complex64))
sino_filtered_freq = tf.multiply(sino_freq,tf.cast(reco_filter,dtype=tf.complex64))
sinogram_filtered = tf.math.real(tf.signal.ifft(sino_filtered_freq))
reco = cone_backprojection3d(sinogram_filtered, geometry)
plt.figure()
plt.imshow(np.squeeze(reco)[volume_shape[0]//2], cmap=plt.get_cmap('gist_gray'))
plt.axis('off')
plt.savefig('3d_cone_reco.png', dpi=150, transparent=False, bbox_inches='tight')
def example_cone_3d():
# ------------------ Declare Parameters ------------------
# Volume Parameters:
volume_size = 256
volume_shape = [volume_size, volume_size, volume_size]
volume_spacing = [0.5, 0.5, 0.5]
# Detector Parameters:
detector_shape = [2 * volume_size, 2 * volume_size]
detector_spacing = [1, 1]
# Trajectory Parameters:
number_of_projections = 360
angular_range = 2 * np.pi
source_detector_distance = 1200
source_isocenter_distance = 750
# create Geometry class
geometry = GeometryCone3D(volume_shape, volume_spacing, detector_shape,
detector_spacing, number_of_projections,
angular_range, source_detector_distance,
source_isocenter_distance)
geometry.set_projection_matrices(
circular_trajectory.circular_trajectory_3d(geometry))
# Get Phantom 3d
phantom = shepp_logan.shepp_logan_3d(volume_shape)
phantom = np.expand_dims(phantom, axis=0)
config = tf.ConfigProto()
config.gpu_options.per_process_gpu_memory_fraction = 0.5
config.gpu_options.allow_growth = True
# ------------------ Call Layers ------------------
with tf.Session(config=config) as sess:
result = cone_projection3d(phantom, geometry)
sinogram = result.eval()
filter = ram_lak_3D(geometry)
sino_freq = np.fft.fft(sinogram, axis=-1)
filtered_sino_freq = sino_freq * filter
filtered_sino = np.fft.ifft(filtered_sino_freq, axis=-1)
result_back_proj = cone_backprojection3d(filtered_sino, geometry)
reco = result_back_proj.eval()
plt.figure()
plt.imshow(np.squeeze(reco)[volume_shape[0] // 2],
cmap=plt.get_cmap('gist_gray'))
plt.axis('off')
plt.savefig('3d_cone_reco.png',
dpi=150,
transparent=False,
bbox_inches='tight')
def example_parallel_3d():
# ------------------ Declare Parameters ------------------
# Volume Parameters:
volume_size = 256
volume_shape = [volume_size, volume_size, volume_size]
volume_spacing = [1, 1, 1]
# Detector Parameters:
detector_size = 350
detector_shape = [detector_size, detector_size]
detector_spacing = [1, 1]
# Trajectory Parameters:
number_of_projections = 180
angular_range = 1 * np.pi
# create Geometry class
geometry = GeometryParallel3D(volume_shape, volume_spacing, detector_shape,
detector_spacing, number_of_projections,
angular_range)
geometry.set_trajectory(circular_trajectory_3d(geometry))
# Get Phantom
phantom = shepp_logan.shepp_logan_3d(volume_shape)
# ------------------ Call Layers ------------------
with tf.Session() as sess:
#result = parallel_projection3d(phantom, geometry)
result = par_projection3d(phantom, geometry)
sinogram = result.eval()
sinogram = sinogram + np.random.normal(loc=np.mean(np.abs(sinogram)),
scale=np.std(sinogram),
size=sinogram.shape) * 0.02
reco_filter = filters.ram_lak_3D(geometry)
sino_freq = np.fft.fft(sinogram, axis=2)
sino_filtered_freq = np.multiply(sino_freq, reco_filter)
sinogram_filtered = np.fft.ifft(sino_filtered_freq, axis=2)
result_back_proj = par_backprojection3d(sinogram_filtered, geometry)
reco = result_back_proj.eval()
import pyconrad as pyc
pyc.setup_pyconrad()
pyc.imshow(phantom)
pyc.imshow(sinogram)
pyc.imshow(reco)
a = 5
def __init__(self, geometry):
self.geometry = geometry
self.cosine_weight = tf.get_variable(
name='cosine_weight',
dtype=tf.float32,
initializer=ct_weights.cosine_weights_3d(self.geometry),
trainable=False)
self.redundancy_weight = tf.get_variable(
name='redundancy_weight',
dtype=tf.float32,
initializer=ct_weights.parker_weights_3d(self.geometry),
trainable=False)
self.filter = tf.get_variable(name='reco_filter',
dtype=tf.float32,
initializer=ram_lak_3D(self.geometry),
trainable=False)
def __init__(self, geometry):
"""
Parameters
----------
geometry : GeometryCone3D
The geometry used for reconstruction
"""
self.geometry = geometry
self.cosine_weight = tf.get_variable(
name='cosine_weight',
dtype=tf.float32,
initializer=ct_weights.cosine_weights_3d(self.geometry),
trainable=False)
self.filter = tf.get_variable(name='reco_filter',
dtype=tf.float32,
initializer=ram_lak_3D(self.geometry),
trainable=False)
