def test_localNetUpSampleResnetBlock():
"""
Test the layer.LocalNetUpSampleResnetBlock class, its default attributes and its call() function.
"""
batch_size = 5
channels = 4
input_size = (32, 32, 16)
output_size = (64, 64, 32)
nonskip_tensor_size = (batch_size, ) + input_size + (channels, )
skip_tensor_size = (batch_size, ) + output_size + (channels, )
# Test __init__() and build()
model = layer.LocalNetUpSampleResnetBlock(8)
model.build([nonskip_tensor_size, skip_tensor_size])
assert model._filters == 8
assert model._use_additive_upsampling is True
assert isinstance(model._deconv3d_block, type(layer.Deconv3dBlock(8)))
assert isinstance(model._additive_upsampling,
type(layer.AdditiveUpSampling(output_size)))
assert isinstance(model._conv3d_block, type(layer.Conv3dBlock(8)))
assert isinstance(model._residual_block,
type(layer.LocalNetResidual3dBlock(8)))
def test_init_LocalNet():
"""
Testing init of LocalNet as expected
"""
local_test = loc.LocalNet(
image_size=[1, 2, 3],
out_channels=3,
num_channel_initial=3,
extract_levels=[1, 2, 3],
out_kernel_initializer="he_normal",
out_activation="softmax",
)
# Asserting initialised var for extract_levels is the same - Pass
assert local_test._extract_levels == [1, 2, 3]
# Asserting initialised var for extract_max_level is the same - Pass
assert local_test._extract_max_level == 3
# Asserting initialised var for extract_min_level is the same - Pass
assert local_test._extract_min_level == 1
# Assert downsample blocks type is correct, Pass
assert all(
isinstance(item, type(layer.DownSampleResnetBlock(12)))
for item in local_test._downsample_blocks
)
# Assert number of downsample blocks is correct (== max level), Pass
assert len(local_test._downsample_blocks) == 3
# Assert upsample blocks type is correct, Pass
assert all(
isinstance(item, type(layer.LocalNetUpSampleResnetBlock(12)))
for item in local_test._upsample_blocks
)
# Assert number of upsample blocks is correct (== max level - min level), Pass
assert len(local_test._upsample_blocks) == 3 - 1
# Assert upsample blocks type is correct, Pass
assert all(
isinstance(item, type(layer.Conv3dWithResize(12, filters=3)))
for item in local_test._extract_layers
)
# Assert number of upsample blocks is correct (== extract_levels), Pass
assert len(local_test._extract_layers) == 3
def __init__(self,
image_size, out_channels,
num_channel_initial, extract_levels,
out_kernel_initializer, out_activation,
**kwargs):
"""
image is encoded gradually, i from level 0 to E
then it is decoded gradually, j from level E to D
some of the decoded level are used for generating extractions
so extract_levels are between [0, E] with E = max(extract_levels) and D = min(extract_levels)
:param out_channels: number of channels for the extractions
:param num_channel_initial:
:param extract_levels:
:param out_kernel_initializer:
:param out_activation:
:param kwargs:
"""
super(LocalNet, self).__init__(**kwargs)
# save parameters
self._extract_levels = extract_levels
self._extract_max_level = max(self._extract_levels) # E
self._extract_min_level = min(self._extract_levels) # D
# init layer variables
nc = [num_channel_initial * (2 ** level) for level in range(self._extract_max_level + 1)] # level 0 to E
self._downsample_blocks = [layer.DownSampleResnetBlock(filters=nc[i], kernel_size=7 if i == 0 else 3)
for i in range(self._extract_max_level)] # level 0 to E-1
self._conv3d_block = layer.Conv3dBlock(filters=nc[-1]) # level E
self._upsample_blocks = [layer.LocalNetUpSampleResnetBlock(nc[level]) for level in
range(self._extract_max_level - 1, self._extract_min_level - 1, -1)] # level D to E-1
self._extract_layers = [
# if kernels are not initialized by zeros, with init NN, extract may be too large
layer.Conv3dWithResize(output_shape=image_size, filters=out_channels,
kernel_initializer=out_kernel_initializer,
activation=out_activation)
for _ in self._extract_levels]
def __init__(
self,
image_size: tuple,
out_channels: int,
num_channel_initial: int,
extract_levels: List[int],
out_kernel_initializer: str,
out_activation: str,
name: str = "LocalNet",
**kwargs,
):
"""
Image is encoded gradually, i from level 0 to E,
then it is decoded gradually, j from level E to D.
Some of the decoded levels are used for generating extractions.
So, extract_levels are between [0, E] with E = max(extract_levels),
and D = min(extract_levels).
:param image_size: such as (dim1, dim2, dim3)
:param out_channels: number of channels for the extractions
:param num_channel_initial: number of initial channels.
:param extract_levels: number of extraction levels.
:param out_kernel_initializer: initializer to use for kernels.
:param out_activation: activation to use at end layer.
:param name: name of the backbone.
:param kwargs: additional arguments.
"""
super().__init__(
image_size=image_size,
out_channels=out_channels,
num_channel_initial=num_channel_initial,
out_kernel_initializer=out_kernel_initializer,
out_activation=out_activation,
name=name,
**kwargs,
)
# save parameters
self._extract_levels = extract_levels
self._extract_max_level = max(self._extract_levels) # E
self._extract_min_level = min(self._extract_levels) # D
# init layer variables
num_channels = [
num_channel_initial * (2**level)
for level in range(self._extract_max_level + 1)
] # level 0 to E
self._downsample_blocks = [
layer.DownSampleResnetBlock(filters=num_channels[i],
kernel_size=7 if i == 0 else 3)
for i in range(self._extract_max_level)
] # level 0 to E-1
self._conv3d_block = layer.Conv3dBlock(
filters=num_channels[-1]) # level E
self._upsample_blocks = [
layer.LocalNetUpSampleResnetBlock(num_channels[level])
for level in range(self._extract_max_level -
1, self._extract_min_level - 1, -1)
] # level D to E-1
self._extract_layers = [
# if kernels are not initialized by zeros, with init NN, extract may be too large
layer.Conv3dWithResize(
output_shape=image_size,
filters=out_channels,
kernel_initializer=out_kernel_initializer,
activation=out_activation,
) for _ in self._extract_levels
]
def __init__(
self,
image_size: tuple,
out_channels: int,
num_channel_initial: int,
extract_levels: List[int],
out_kernel_initializer: str,
out_activation: str,
control_points: (tuple, None) = None,
**kwargs,
):
"""
Image is encoded gradually, i from level 0 to E,
then it is decoded gradually, j from level E to D.
Some of the decoded levels are used for generating extractions.
So, extract_levels are between [0, E] with E = max(extract_levels),
and D = min(extract_levels).
:param image_size: tuple, such as (dim1, dim2, dim3)
:param out_channels: int, number of channels for the extractions
:param num_channel_initial: int, number of initial channels.
:param extract_levels: list of int, number of extraction levels.
:param out_kernel_initializer: str, initializer to use for kernels.
:param out_activation: str, activation to use at end layer.
:param control_points: (tuple, None), specify the distance between control points (in voxels).
:param kwargs:
"""
super(LocalNet, self).__init__(**kwargs)
# save parameters
self._extract_levels = extract_levels
self._extract_max_level = max(self._extract_levels) # E
self._extract_min_level = min(self._extract_levels) # D
# init layer variables
num_channels = [
num_channel_initial * (2**level)
for level in range(self._extract_max_level + 1)
] # level 0 to E
self._downsample_blocks = [
layer.DownSampleResnetBlock(filters=num_channels[i],
kernel_size=7 if i == 0 else 3)
for i in range(self._extract_max_level)
] # level 0 to E-1
self._conv3d_block = layer.Conv3dBlock(
filters=num_channels[-1]) # level E
self._upsample_blocks = [
layer.LocalNetUpSampleResnetBlock(num_channels[level])
for level in range(self._extract_max_level -
1, self._extract_min_level - 1, -1)
] # level D to E-1
self._extract_layers = [
# if kernels are not initialized by zeros, with init NN, extract may be too large
layer.Conv3dWithResize(
output_shape=image_size,
filters=out_channels,
kernel_initializer=out_kernel_initializer,
activation=out_activation,
) for _ in self._extract_levels
]
self.resize = (layer.ResizeCPTransform(control_points)
if control_points is not None else False)
self.interpolate = (layer.BSplines3DTransform(control_points,
image_size)
if control_points is not None else False)