def conv_t_block(size,
input_1,
input_2,
kernel,
stride,
activation,
kinit,
padding,
axis,
batch_normalization=False):
conv1 = Convolution2DTranspose(256, (2, 2),
strides=(2, 2),
padding=padding)(input_1)
conv1 = BatchNormalization()(conv1)
if input_2 is not None:
conv1 = concatenate([conv1, input_2], axis=axis)
conv2 = Convolution2D(256,
kernel_size=kernel,
strides=stride,
activation=activation,
kernel_initializer=kinit,
padding=padding)(conv1)
conv2 = BatchNormalization()(conv2) if batch_normalization else conv2
conv3 = Convolution2D(256,
kernel_size=kernel,
strides=stride,
activation=activation,
kernel_initializer=kinit,
padding=padding)(conv2)
return conv3
def f(input):
layer = Convolution2DTranspose(nb_filter, kernel_size=(rk, ck), strides=(2,2),\
padding="same", use_bias=False,\
kernel_initializer="Orthogonal",name='conv-'+name)(input)
layer_norm = BatchNormalization(name='BN-'+name)(layer)
layer_norm_relu = Activation(activation = "relu",name='Relu-'+name)(layer_norm)
if boolean_dropout:
layer_norm_relu = Dropout(0.1) (layer_norm_relu)
return layer_norm_relu
def conv_t_block(size,
input_1,
input_2,
kernel,
stride,
activation,
kinit,
padding,
axis,
batch_normalization=False,
dropout=None):
use_bias = not batch_normalization
conv1 = Convolution2DTranspose(256, (2, 2),
strides=(2, 2),
padding=padding,
use_bias=use_bias)(input_1)
conv1 = BatchNormalization()(conv1)
conv1 = Activation(activation)(conv1)
if input_2 is not None:
input_2 = BatchNormalization()(input_2)
conv1 = concatenate([conv1, input_2], axis=axis)
conv2 = Convolution2D(256,
kernel_size=kernel,
strides=stride,
kernel_initializer=kinit,
padding=padding,
use_bias=use_bias)(conv1)
conv2 = BatchNormalization()(conv2) if batch_normalization else conv2
conv2 = Activation(activation)(conv2)
if dropout is not None:
conv2 = Dropout(rate=dropout)(conv2)
conv3 = Convolution2D(256,
kernel_size=kernel,
strides=stride,
kernel_initializer=kinit,
padding=padding,
use_bias=False)(conv2)
conv3 = BatchNormalization()(conv3) if batch_normalization else conv2
conv3 = Activation(activation)(conv3)
return conv3
def up_conv(filters, kernel_size, stride):
return Convolution2DTranspose(filters, kernel_size, stride)
def unet_v14(
name,
input_shapes,
output_shapes,
kernel=3,
stride=1,
activation='elu',
output_channels=2,
kinit='RandomUniform',
batch_norm=True,
padding='same',
axis=3,
crop=0,
mpadd=0,
):
nr_classes = output_channels
if len(input_shapes["input_1"]) == 3:
input_1_height, input_1_width, input_1_channels = input_shapes[
"input_1"]
else:
timestamps_1, input_1_height, input_1_width, input_1_channels = input_shapes[
"input_1"]
inputs = Input((input_1_height, input_1_width, input_1_channels))
conv1 = ZeroPadding2D((crop, crop))(inputs)
conv1 = Convolution2D(32,
kernel_size=kernel,
strides=stride,
activation=activation,
kernel_initializer=kinit,
padding=padding)(conv1)
conv1 = BatchNormalization()(conv1)
conv1 = Convolution2D(32,
kernel_size=kernel,
strides=stride,
activation=activation,
kernel_initializer=kinit,
padding=padding)(conv1)
conv1 = BatchNormalization()(conv1)
pool1 = MaxPooling2D(pool_size=(2, 2))(conv1)
conv2 = Convolution2D(64,
kernel_size=kernel,
strides=stride,
activation=activation,
kernel_initializer=kinit,
padding=padding)(pool1)
conv2 = BatchNormalization()(conv2)
conv2 = Convolution2D(64,
kernel_size=kernel,
strides=stride,
activation=activation,
kernel_initializer=kinit,
padding=padding)(conv2)
conv2 = BatchNormalization()(conv2)
pool2 = MaxPooling2D(pool_size=(2, 2))(conv2)
conv3 = Convolution2D(128,
kernel_size=kernel,
strides=stride,
activation=activation,
kernel_initializer=kinit,
padding=padding)(pool2)
conv3 = BatchNormalization()(conv3)
conv3 = Convolution2D(128,
kernel_size=kernel,
strides=stride,
activation=activation,
kernel_initializer=kinit,
padding=padding)(conv3)
conv3 = BatchNormalization()(conv3)
pool3 = MaxPooling2D(pool_size=(2, 2))(conv3)
conv4 = Convolution2D(256,
kernel_size=kernel,
strides=stride,
activation=activation,
kernel_initializer=kinit,
padding=padding)(pool3)
conv4 = BatchNormalization()(conv4)
conv4 = Convolution2D(256,
kernel_size=kernel,
strides=stride,
activation=activation,
kernel_initializer=kinit,
padding=padding)(conv4)
conv4 = BatchNormalization()(conv4)
pool4 = MaxPooling2D(pool_size=(2, 2))(conv4)
conv5 = Convolution2D(512,
kernel_size=kernel,
strides=stride,
activation=activation,
kernel_initializer=kinit,
padding=padding)(pool4)
conv5 = BatchNormalization()(conv5)
conv5 = Convolution2D(512,
kernel_size=kernel,
strides=stride,
activation=activation,
kernel_initializer=kinit,
padding=padding)(conv5)
conv5 = BatchNormalization()(conv5)
up6 = concatenate([
Convolution2DTranspose(256, (2, 2), strides=(2, 2),
padding=padding)(conv5), conv4
],
axis=axis)
conv6 = Convolution2D(256,
kernel_size=kernel,
strides=stride,
activation=activation,
kernel_initializer=kinit,
padding=padding)(up6)
conv6 = BatchNormalization()(conv6) if batch_norm else conv6
conv6 = Convolution2D(256,
kernel_size=kernel,
strides=stride,
activation=activation,
kernel_initializer=kinit,
padding=padding)(conv6)
conv6 = BatchNormalization()(conv6) if batch_norm else conv6
up7 = concatenate([
Convolution2DTranspose(128, (2, 2), strides=(2, 2),
padding=padding)(conv6), conv3
],
axis=axis)
conv7 = Convolution2D(128,
kernel_size=kernel,
strides=stride,
activation=activation,
kernel_initializer=kinit,
padding=padding)(up7)
conv7 = BatchNormalization()(conv7) if batch_norm else conv7
conv7 = Convolution2D(128,
kernel_size=kernel,
strides=stride,
activation=activation,
kernel_initializer=kinit,
padding=padding)(conv7)
conv7 = BatchNormalization()(conv7) if batch_norm else conv7
up8 = concatenate([
Convolution2DTranspose(64, (2, 2), strides=(2, 2),
padding=padding)(conv7), conv2
],
axis=axis)
conv8 = Convolution2D(64,
kernel_size=kernel,
strides=stride,
activation=activation,
kernel_initializer=kinit,
padding=padding)(up8)
conv8 = BatchNormalization()(conv8) if batch_norm else conv8
conv8 = Convolution2D(64,
kernel_size=kernel,
strides=stride,
activation=activation,
kernel_initializer=kinit,
padding=padding)(conv8)
conv8 = BatchNormalization()(conv8) if batch_norm else conv8
up9 = concatenate([
Convolution2DTranspose(32, (2, 2), strides=(2, 2),
padding=padding)(conv8), conv1
],
axis=axis)
conv9 = Convolution2D(32,
kernel_size=kernel,
strides=stride,
activation=activation,
kernel_initializer=kinit,
padding=padding)(up9)
conv9 = BatchNormalization()(conv9) if batch_norm else conv9
conv9 = Convolution2D(32,
kernel_size=kernel,
strides=stride,
activation=activation,
kernel_initializer=kinit,
padding=padding)(conv9)
conv9 = BatchNormalization()(conv9) if batch_norm else conv9
conv9 = Cropping2D((mpadd, mpadd))(conv9)
conv10 = Convolution2D(nr_classes, (1, 1),
activation='softmax',
name="output_1")(conv9)
model = Model(inputs=[inputs], outputs=[conv10])
return model
def Net(n_classes, input):
# ---------left branch -----
x = conv_block(input, 32, (3, 3), strides=1, name='L_conv1-1')
L1 = conv_block(x, 32, (3, 3), strides=1, name='L_conv1-2')
x = MaxPooling2D((2, 2), strides=(2, 2), padding='same')(L1)
# 256 -> 128
x = conv_block(x, 64, (3, 3), strides=1, name='L_conv2-1')
L2 = conv_block(x, 64, (3, 3), strides=1, name='L_conv2-2')
x = MaxPooling2D((2, 2), strides=(2, 2), padding='same')(L2)
# 128 -> 64
x = conv_block(x, 128, (3, 3), strides=1, name='L_conv3-1')
L3 = conv_block(x, 128, (3, 3), strides=1, name='L_conv3-2')
x = MaxPooling2D((2, 2), strides=(2, 2), padding='same')(L3)
# 64 -> 32
x = conv_block(x, 256, (3, 3), strides=1, name='L_conv4-1')
L4 = conv_block(x, 256, (3, 3), strides=1, name='L_conv4-2')
x = MaxPooling2D((2, 2), strides=(2, 2), padding='same')(L4)
# 32 -> 16
x = conv_block(x, 512, (3, 3), strides=1, name='bottom-1')
x = conv_block(x, 512, (3, 3), strides=1, dila=2, name='bottom-2')
L5 = conv_block(x, 512, (3, 3), strides=1, name='bottom-3')
# 16
# ---------Right branch -----
# 16 -> 32
x = Convolution2DTranspose(
256, kernel_size=2, strides=2, padding='same', name='R_conv1-1')(L5)
x = BatchNormalization(name='R_conv1-1_' + 'bn')(x)
x = conv_block(Concatenate(axis=-1)([x, L4]), 256, (3, 3), strides=1, name='R_conv1-2')
x = conv_block(x, 256, (3, 3), strides=1, name='R_conv1-3')
# 32 -> 64
x = Convolution2DTranspose(
128, kernel_size=2, strides=2, padding='same', name='R_conv2-1')(x)
x = BatchNormalization(name='R_conv2-1_' + 'bn')(x)
x = conv_block(Concatenate(axis=-1)([x, L3]), 128, (3, 3), strides=1, name='R_conv2-2')
x = conv_block(x, 128, (3, 3), strides=1, name='R_conv2-3')
# 64 -> 128
x = Convolution2DTranspose(
64, kernel_size=2, strides=2, padding='same', name='R_conv3-1')(x)
x = BatchNormalization(name='R_conv3-1_' + 'bn')(x)
x = conv_block(Concatenate(axis=-1) ([x, L2]), 64, (3, 3), strides=1, name='R_conv3-2')
x = conv_block(x, 64, (3, 3), strides=1, name='R_conv3-3')
# 128 -> 256
x = Convolution2DTranspose(
32, kernel_size=2, strides=2, padding='same', name='R_conv4-1')(x)
x = BatchNormalization(name='R_conv4-1_' + 'bn')(x)
x = conv_block(Concatenate(axis=-1) ([x, L1]), 32, (3, 3), strides=1, name='R_conv4-2')
x = conv_block(x, 32, (3, 3), strides=1, name='R_conv4-3')
final = Conv2D(n_classes, (1, 1), name='final_out')(x)
final = Activation('softmax', name='softmax_1')(final)
return final
def build(self):
self.inputs = Input(shape=(256, 256, 3), name="img")
self.conv_1 = Conv2D(32,
3,
strides=(2, 2),
use_bias=True,
padding='same',
name="conv_1")(self.inputs)
self.relu_1 = ReLU()(self.conv_1)
self.residualb_1 = self.residual_block(self.relu_1, 32, "residualb_1")
self.residualb_2 = self.residual_block(self.residualb_1, 32,
"residualb_2")
self.residualb_3 = self.residual_block(self.residualb_2, 32,
"residualb_3")
self.residualb_4 = self.residual_block(self.residualb_3, 32,
"residualb_4")
self.residualb_5 = self.residual_block(self.residualb_4, 32,
"residualb_5")
self.residualb_6 = self.residual_block(self.residualb_5, 32,
"residualb_6")
self.residualb_7 = self.residual_block(self.residualb_6, 32,
"residualb_7")
self.residualb_8 = self.residual_block_id(self.residualb_7, 64,
"residualb_8")
self.residualb_9 = self.residual_block(self.residualb_8, 64,
"residualb_9")
self.residualb_10 = self.residual_block(self.residualb_9, 64,
"residualb_10")
self.residualb_11 = self.residual_block(self.residualb_10, 64,
"residualb_11")
self.residualb_12 = self.residual_block(self.residualb_11, 64,
"residualb_12")
self.residualb_13 = self.residual_block(self.residualb_12, 64,
"residualb_13")
self.residualb_14 = self.residual_block(self.residualb_13, 64,
"residualb_14")
self.residualb_15 = self.residual_block(self.residualb_14, 64,
"residualb_15")
self.residualb_16 = self.residual_block_id(self.residualb_15, 128,
"residualb_16")
self.residualb_17 = self.residual_block(self.residualb_16, 128,
"residualb_17")
self.residualb_18 = self.residual_block(self.residualb_17, 128,
"residualb_18")
self.residualb_19 = self.residual_block(self.residualb_18, 128,
"residualb_19")
self.residualb_20 = self.residual_block(self.residualb_19, 128,
"residualb_20")
self.residualb_21 = self.residual_block(self.residualb_20, 128,
"residualb_21")
self.residualb_22 = self.residual_block(self.residualb_21, 128,
"residualb_22")
self.residualb_23 = self.residual_block(self.residualb_22, 128,
"residualb_23")
self.residualb_24 = self.residual_block_id(self.residualb_23, 256,
"residualb_24")
self.residualb_25 = self.residual_block(self.residualb_24, 256,
"residualb_25")
self.residualb_26 = self.residual_block(self.residualb_25, 256,
"residualb_26")
self.residualb_27 = self.residual_block(self.residualb_26, 256,
"residualb_27")
self.residualb_28 = self.residual_block(self.residualb_27, 256,
"residualb_28")
self.residualb_29 = self.residual_block(self.residualb_28, 256,
"residualb_29")
self.residualb_30 = self.residual_block(self.residualb_29, 256,
"residualb_30")
self.residualb_31 = self.residual_block(self.residualb_30, 256,
"residualb_31")
self.max_pool_1 = MaxPool2D(pool_size=(3, 3),
strides=(2, 2),
padding='same')(self.residualb_31)
self.dconv_1 = DepthwiseConv2D(kernel_size=(3, 3),
strides=(2, 2),
padding='same',
name="dconv_1")(self.residualb_31)
self.conv_2 = Conv2D(filters=256,
kernel_size=(1, 1),
strides=(1, 1),
padding='valid',
name="conv_2")(self.dconv_1)
self.add_1 = Add()([self.conv_2, self.max_pool_1])
self.relu_2 = ReLU()(self.add_1)
self.residualb_32 = self.residual_block(self.relu_2, 256,
"residualb_32")
self.residualb_33 = self.residual_block(self.residualb_32, 256,
"residualb_33")
self.residualb_34 = self.residual_block(self.residualb_33, 256,
"residualb_34")
self.residualb_35 = self.residual_block(self.residualb_34, 256,
"residualb_35")
self.residualb_36 = self.residual_block(self.residualb_35, 256,
"residualb_36")
self.residualb_37 = self.residual_block(self.residualb_36, 256,
"residualb_37")
#split key
self.residualb_38 = self.residual_block(self.residualb_37, 256,
"residualb_38")
#BRANCH 1
self.conv_transpose_1 = Convolution2DTranspose(filters=256,
kernel_size=(2, 2),
strides=(2, 2),
name="convt_1")(
self.residualb_38)
self.relu_3 = ReLU()(self.conv_transpose_1)
self.add_2 = Add()([self.residualb_31, self.relu_3])
#split key
self.residualb_39 = self.residual_block(self.add_2, 256,
"residualb_39")
self.conv_transpose_2 = Convolution2DTranspose(filters=128,
kernel_size=(2, 2),
strides=(2, 2),
name="convt_2")(
self.residualb_39)
self.relu_4 = ReLU()(self.conv_transpose_2)
self.add_3 = Add()([self.residualb_23, self.relu_4])
#split key
self.residualb_40 = self.residual_block(self.add_3, 128,
"residualb_40")
# output block 1
self.conv_3 = Conv2D(filters=2,
kernel_size=(1, 1),
strides=(1, 1),
padding='same',
name="conv_3")(self.residualb_40)
#self.reshape_1 = tf.reshape(self.conv_3,[1,2048,1])
self.reshape_1 = Reshape([-1, 1])(self.conv_3)
self.conv_4 = Conv2D(filters=2,
kernel_size=(1, 1),
strides=(1, 1),
name="conv_4")(self.residualb_39)
#self.reshape_2 = tf.reshape(self.conv_4,[1,512,1])
self.reshape_2 = Reshape([-1, 1])(self.conv_4)
self.conv_5 = Conv2D(filters=6,
kernel_size=(1, 1),
strides=(1, 1),
name="conv_5")(self.residualb_38)
#self.reshape_3 = tf.reshape(self.conv_5, [1,384,1])
self.reshape_3 = Reshape([-1, 1])(self.conv_5)
self.concat_1 = Concatenate(axis=1)(
[self.reshape_1, self.reshape_2, self.reshape_3])
#output block 2
self.conv_6 = Conv2D(filters=36,
kernel_size=(1, 1),
strides=(1, 1),
padding='same',
name="conv_6")(self.residualb_40)
#self.reshape_4 = tf.reshape(self.conv_6,[1,2048,18])
self.reshape_4 = Reshape([-1, 18])(self.conv_6)
self.conv_7 = Conv2D(filters=36,
kernel_size=(1, 1),
strides=(1, 1),
padding='same',
name="conv_7")(self.residualb_39)
#self.reshape_5 = tf.reshape(self.conv_7,[1,512,18])
self.reshape_5 = Reshape([-1, 18])(self.conv_7)
self.conv_8 = Conv2D(filters=108,
kernel_size=(1, 1),
strides=(1, 1),
name="conv_8")(self.residualb_38)
#self.reshape_6 = tf.reshape(self.conv_8,[1,384,18])
self.reshape_6 = Reshape([-1, 18])(self.conv_8)
self.concat_2 = Concatenate(axis=1)(
[self.reshape_4, self.reshape_5, self.reshape_6])
self.init_model()
return self.concat_1, self.concat_2