def coarsemask_decoder_net(
self,
images,
is_training=None,
norm_activation=nn_ops.norm_activation_builder()):
"""Coarse mask decoder network architecture.
Args:
images: A tensor of size [batch, height_in, width_in, channels_in].
is_training: Whether batch_norm layers are in training mode.
norm_activation: an operation that includes a batch normalization layer
followed by a relu layer(optional).
Returns:
images: A feature tensor of size [batch, output_size, output_size,
num_channels]
"""
for i in range(self._num_convs):
images = tf.keras.layers.Conv2D(
self._num_downsample_channels,
kernel_size=(3, 3),
bias_initializer=tf.zeros_initializer(),
kernel_initializer=tf.keras.initializers.RandomNormal(
stddev=0.01),
activation=None,
padding='same',
name='coarse-class-%d' % i)(images)
images = norm_activation(name='coarse-class-%d-bn' % i)(
images, is_training=is_training)
return images
def __init__(self,
num_classes,
num_downsample_channels,
mask_crop_size,
num_convs,
coarse_mask_thr,
gt_upsample_scale,
norm_activation=nn_ops.norm_activation_builder()):
"""Initialize params to build ShapeMask coarse and fine prediction head.
Args:
num_classes: `int` number of mask classification categories.
num_downsample_channels: `int` number of filters at mask head.
mask_crop_size: feature crop size.
num_convs: `int` number of stacked convolution before the last prediction
layer.
coarse_mask_thr: the threshold for suppressing noisy coarse prediction.
gt_upsample_scale: scale for upsampling groundtruths.
norm_activation: an operation that includes a batch normalization layer
followed by a relu layer(optional).
"""
self._mask_num_classes = num_classes
self._num_downsample_channels = num_downsample_channels
self._mask_crop_size = mask_crop_size
self._num_convs = num_convs
self._coarse_mask_thr = coarse_mask_thr
self._gt_upsample_scale = gt_upsample_scale
self._class_predict_conv = tf.keras.layers.Conv2D(
self._mask_num_classes,
kernel_size=(1, 1),
# Focal loss bias initialization to have foreground 0.01 probability.
bias_initializer=tf.constant_initializer(-np.log((1 - 0.01) /
0.01)),
kernel_initializer=tf.keras.initializers.RandomNormal(mean=0,
stddev=0.01),
padding='same',
name='affinity-class-predict')
self._upsample_conv = tf.keras.layers.Conv2DTranspose(
self._num_downsample_channels // 2,
(self._gt_upsample_scale, self._gt_upsample_scale),
(self._gt_upsample_scale, self._gt_upsample_scale))
self._fine_class_conv = []
self._fine_class_bn = []
for i in range(self._num_convs):
self._fine_class_conv.append(
tf.keras.layers.Conv2D(
self._num_downsample_channels,
kernel_size=(3, 3),
bias_initializer=tf.zeros_initializer(),
kernel_initializer=tf.keras.initializers.RandomNormal(
stddev=0.01),
activation=None,
padding='same',
name='fine-class-%d' % i))
self._fine_class_bn.append(
norm_activation(name='fine-class-%d-bn' % i))
def __init__(self,
num_classes,
num_downsample_channels,
mask_crop_size,
use_category_for_mask,
num_convs,
norm_activation=nn_ops.norm_activation_builder()):
"""Initialize params to build ShapeMask coarse and fine prediction head.
Args:
num_classes: `int` number of mask classification categories.
num_downsample_channels: `int` number of filters at mask head.
mask_crop_size: feature crop size.
use_category_for_mask: use class information in mask branch.
num_convs: `int` number of stacked convolution before the last prediction
layer.
norm_activation: an operation that includes a normalization layer
followed by an optional activation layer.
"""
self._mask_num_classes = num_classes if use_category_for_mask else 1
self._use_category_for_mask = use_category_for_mask
self._num_downsample_channels = num_downsample_channels
self._mask_crop_size = mask_crop_size
self._num_convs = num_convs
self._norm_activation = norm_activation
self._coarse_mask_fc = tf.keras.layers.Dense(
self._num_downsample_channels, name='coarse-mask-fc')
self._class_conv = []
self._class_norm_activation = []
for i in range(self._num_convs):
self._class_conv.append(
tf.keras.layers.Conv2D(
self._num_downsample_channels,
kernel_size=(3, 3),
bias_initializer=tf.zeros_initializer(),
kernel_initializer=tf.keras.initializers.RandomNormal(
stddev=0.01),
padding='same',
name='coarse-mask-class-%d' % i))
self._class_norm_activation.append(
norm_activation(name='coarse-mask-class-%d-bn' % i))
self._class_predict = tf.keras.layers.Conv2D(
self._mask_num_classes,
kernel_size=(1, 1),
# Focal loss bias initialization to have foreground 0.01 probability.
bias_initializer=tf.constant_initializer(-np.log((1 - 0.01) /
0.01)),
kernel_initializer=tf.keras.initializers.RandomNormal(stddev=0.01),
padding='same',
name='coarse-mask-class-predict')
def __init__(self,
resnet_depth,
activation='relu',
norm_activation=nn_ops.norm_activation_builder(
activation='relu'),
data_format='channels_last'):
"""ResNet initialization function.
Args:
resnet_depth: `int` depth of ResNet backbone model.
norm_activation: an operation that includes a normalization layer
followed by an optional activation layer.
data_format: `str` either "channels_first" for `[batch, channels, height,
width]` or "channels_last for `[batch, height, width, channels]`.
"""
self._resnet_depth = resnet_depth
if activation == 'relu':
self._activation_op = tf.nn.relu
elif activation == 'swish':
self._activation_op = tf.nn.swish
else:
raise ValueError('Unsupported activation `{}`.'.format(activation))
self._norm_activation = norm_activation
self._data_format = data_format
model_params = {
10: {'block': self.residual_block, 'layers': [1, 1, 1, 1]},
18: {'block': self.residual_block, 'layers': [2, 2, 2, 2]},
34: {'block': self.residual_block, 'layers': [3, 4, 6, 3]},
50: {'block': self.bottleneck_block, 'layers': [3, 4, 6, 3]},
101: {'block': self.bottleneck_block, 'layers': [3, 4, 23, 3]},
152: {'block': self.bottleneck_block, 'layers': [3, 8, 36, 3]},
200: {'block': self.bottleneck_block, 'layers': [3, 24, 36, 3]}
}
if resnet_depth not in model_params:
valid_resnet_depths = ', '.join(
[str(depth) for depth in sorted(model_params.keys())])
raise ValueError(
'The resnet_depth should be in [%s]. Not a valid resnet_depth:'%(
valid_resnet_depths), self._resnet_depth)
params = model_params[resnet_depth]
self._resnet_fn = self.resnet_v1_generator(
params['block'], params['layers'])
def __init__(
self,
min_level,
max_level,
num_classes,
anchors_per_location,
num_convs=4,
num_filters=256,
use_separable_conv=False,
norm_activation=nn_ops.norm_activation_builder(activation='relu')):
"""Initialize params to build RetinaNet head.
Args:
min_level: `int` number of minimum feature level.
max_level: `int` number of maximum feature level.
num_classes: `int` number of classification categories.
anchors_per_location: `int` number of anchors per pixel location.
num_convs: `int` number of stacked convolution before the last prediction
layer.
num_filters: `int` number of filters used in the head architecture.
use_separable_conv: `bool` to indicate whether to use separable
convoluation.
norm_activation: an operation that includes a normalization layer followed
by an optional activation layer.
"""
self._min_level = min_level
self._max_level = max_level
self._num_classes = num_classes
self._anchors_per_location = anchors_per_location
self._num_convs = num_convs
self._num_filters = num_filters
self._use_separable_conv = use_separable_conv
with tf.name_scope('class_net') as scope_name:
self._class_name_scope = tf.name_scope(scope_name)
with tf.name_scope('box_net') as scope_name:
self._box_name_scope = tf.name_scope(scope_name)
self._build_class_net_layers(norm_activation)
self._build_box_net_layers(norm_activation)
def norm_activation_generator(params):
return nn_ops.norm_activation_builder(
momentum=params.batch_norm_momentum,
epsilon=params.batch_norm_epsilon,
trainable=params.batch_norm_trainable,
activation=params.activation)
def __init__(
self,
num_classes,
mask_target_size,
num_convs=4,
num_filters=256,
use_separable_conv=False,
activation='relu',
use_batch_norm=True,
norm_activation=nn_ops.norm_activation_builder(activation='relu')):
"""Initialize params to build Fast R-CNN head.
Args:
num_classes: a integer for the number of classes.
mask_target_size: a integer that is the resolution of masks.
num_convs: `int` number that represents the number of the intermediate
conv layers before the prediction.
num_filters: `int` number that represents the number of filters of the
intermediate conv layers.
use_separable_conv: `bool`, indicating whether the separable conv layers
is used.
activation: activation function. Support 'relu' and 'swish'.
use_batch_norm: 'bool', indicating whether batchnorm layers are added.
norm_activation: an operation that includes a normalization layer followed
by an optional activation layer.
"""
super(MaskrcnnHead, self).__init__(autocast=False)
self._num_classes = num_classes
self._mask_target_size = mask_target_size
self._num_convs = num_convs
self._num_filters = num_filters
if use_separable_conv:
self._conv2d_op = functools.partial(
tf.keras.layers.SeparableConv2D,
depth_multiplier=1,
bias_initializer=tf.zeros_initializer())
else:
self._conv2d_op = functools.partial(
tf.keras.layers.Conv2D,
kernel_initializer=tf.keras.initializers.VarianceScaling(
scale=2, mode='fan_out', distribution='untruncated_normal'),
bias_initializer=tf.zeros_initializer())
if activation == 'relu':
self._activation_op = tf.nn.relu
elif activation == 'swish':
self._activation_op = tf.nn.swish
else:
raise ValueError('Unsupported activation `{}`.'.format(activation))
self._use_batch_norm = use_batch_norm
self._norm_activation = norm_activation
self._conv2d_ops = []
for i in range(self._num_convs):
self._conv2d_ops.append(
self._conv2d_op(
self._num_filters,
kernel_size=(3, 3),
strides=(1, 1),
padding='same',
dilation_rate=(1, 1),
activation=(None
if self._use_batch_norm else self._activation_op),
name='mask-conv-l%d' % i))
self._mask_conv_transpose = tf.keras.layers.Conv2DTranspose(
self._num_filters,
kernel_size=(2, 2),
strides=(2, 2),
padding='valid',
activation=(None if self._use_batch_norm else self._activation_op),
kernel_initializer=tf.keras.initializers.VarianceScaling(
scale=2, mode='fan_out', distribution='untruncated_normal'),
bias_initializer=tf.zeros_initializer(),
name='conv5-mask')
with tf.name_scope('mask_head'):
self._mask_conv2d_op = self._conv2d_op(
self._num_classes,
kernel_size=(1, 1),
strides=(1, 1),
padding='valid',
name='mask_fcn_logits')
def __init__(
self,
num_classes,
num_convs=0,
num_filters=256,
use_separable_conv=False,
num_fcs=2,
fc_dims=1024,
activation='relu',
use_batch_norm=True,
norm_activation=nn_ops.norm_activation_builder(activation='relu')):
"""Initialize params to build OLN box head.
Args:
num_classes: a integer for the number of classes.
num_convs: `int` number that represents the number of the intermediate
conv layers before the FC layers.
num_filters: `int` number that represents the number of filters of the
intermediate conv layers.
use_separable_conv: `bool`, indicating whether the separable conv layers
is used.
num_fcs: `int` number that represents the number of FC layers before the
predictions.
fc_dims: `int` number that represents the number of dimension of the FC
layers.
activation: activation function. Support 'relu' and 'swish'.
use_batch_norm: 'bool', indicating whether batchnorm layers are added.
norm_activation: an operation that includes a normalization layer followed
by an optional activation layer.
"""
self._num_classes = num_classes
self._num_convs = num_convs
self._num_filters = num_filters
if use_separable_conv:
self._conv2d_op = functools.partial(
tf.keras.layers.SeparableConv2D,
depth_multiplier=1,
bias_initializer=tf.zeros_initializer())
else:
self._conv2d_op = functools.partial(
tf.keras.layers.Conv2D,
kernel_initializer=tf.keras.initializers.VarianceScaling(
scale=2, mode='fan_out', distribution='untruncated_normal'),
bias_initializer=tf.zeros_initializer())
self._num_fcs = num_fcs
self._fc_dims = fc_dims
if activation == 'relu':
self._activation_op = tf.nn.relu
elif activation == 'swish':
self._activation_op = tf.nn.swish
else:
raise ValueError('Unsupported activation `{}`.'.format(activation))
self._use_batch_norm = use_batch_norm
self._norm_activation = norm_activation
self._conv_ops = []
self._conv_bn_ops = []
for i in range(self._num_convs):
self._conv_ops.append(
self._conv2d_op(
self._num_filters,
kernel_size=(3, 3),
strides=(1, 1),
padding='same',
dilation_rate=(1, 1),
activation=(None
if self._use_batch_norm else self._activation_op),
name='conv_{}'.format(i)))
if self._use_batch_norm:
self._conv_bn_ops.append(self._norm_activation())
self._fc_ops = []
self._fc_bn_ops = []
for i in range(self._num_fcs):
self._fc_ops.append(
tf.keras.layers.Dense(
units=self._fc_dims,
activation=(None
if self._use_batch_norm else self._activation_op),
name='fc{}'.format(i)))
if self._use_batch_norm:
self._fc_bn_ops.append(self._norm_activation(fused=False))
self._class_predict = tf.keras.layers.Dense(
self._num_classes,
kernel_initializer=tf.keras.initializers.RandomNormal(stddev=0.01),
bias_initializer=tf.zeros_initializer(),
name='class-predict')
self._box_predict = tf.keras.layers.Dense(
self._num_classes * 4,
kernel_initializer=tf.keras.initializers.RandomNormal(stddev=0.001),
bias_initializer=tf.zeros_initializer(),
name='box-predict')
self._score_predict = tf.keras.layers.Dense(
1,
kernel_initializer=tf.keras.initializers.RandomNormal(stddev=0.01),
bias_initializer=tf.zeros_initializer(),
name='score-predict')
def __init__(
self,
min_level,
max_level,
anchors_per_location,
num_convs=2,
num_filters=256,
use_separable_conv=False,
activation='relu',
use_batch_norm=True,
norm_activation=nn_ops.norm_activation_builder(activation='relu')):
"""Initialize params to build Region Proposal Network head.
Args:
min_level: `int` number of minimum feature level.
max_level: `int` number of maximum feature level.
anchors_per_location: `int` number of number of anchors per pixel
location.
num_convs: `int` number that represents the number of the intermediate
conv layers before the prediction.
num_filters: `int` number that represents the number of filters of the
intermediate conv layers.
use_separable_conv: `bool`, indicating whether the separable conv layers
is used.
activation: activation function. Support 'relu' and 'swish'.
use_batch_norm: 'bool', indicating whether batchnorm layers are added.
norm_activation: an operation that includes a normalization layer followed
by an optional activation layer.
"""
super().__init__(autocast=False)
self._min_level = min_level
self._max_level = max_level
self._anchors_per_location = anchors_per_location
if activation == 'relu':
self._activation_op = tf.nn.relu
elif activation == 'swish':
self._activation_op = tf.nn.swish
else:
raise ValueError('Unsupported activation `{}`.'.format(activation))
self._use_batch_norm = use_batch_norm
if use_separable_conv:
self._conv2d_op = functools.partial(
tf.keras.layers.SeparableConv2D,
depth_multiplier=1,
bias_initializer=tf.zeros_initializer())
else:
self._conv2d_op = functools.partial(
tf.keras.layers.Conv2D,
kernel_initializer=tf.keras.initializers.RandomNormal(stddev=0.01),
bias_initializer=tf.zeros_initializer())
self._rpn_conv = self._conv2d_op(
num_filters,
kernel_size=(3, 3),
strides=(1, 1),
activation=(None if self._use_batch_norm else self._activation_op),
padding='same',
name='rpn')
self._rpn_class_conv = self._conv2d_op(
anchors_per_location,
kernel_size=(1, 1),
strides=(1, 1),
padding='valid',
name='rpn-class')
self._rpn_box_conv = self._conv2d_op(
4 * anchors_per_location,
kernel_size=(1, 1),
strides=(1, 1),
padding='valid',
name='rpn-box')
self._norm_activations = {}
if self._use_batch_norm:
for level in range(self._min_level, self._max_level + 1):
self._norm_activations[level] = norm_activation(name='rpn-l%d-bn' %
level)
def __init__(
self,
min_level=3,
max_level=7,
fpn_feat_dims=256,
use_separable_conv=False,
activation='relu',
use_batch_norm=True,
norm_activation=nn_ops.norm_activation_builder(activation='relu')):
"""FPN initialization function.
Args:
min_level: `int` minimum level in FPN output feature maps.
max_level: `int` maximum level in FPN output feature maps.
fpn_feat_dims: `int` number of filters in FPN layers.
use_separable_conv: `bool`, if True use separable convolution for
convolution in FPN layers.
use_batch_norm: 'bool', indicating whether batchnorm layers are added.
norm_activation: an operation that includes a normalization layer
followed by an optional activation layer.
"""
self._min_level = min_level
self._max_level = max_level
self._fpn_feat_dims = fpn_feat_dims
if use_separable_conv:
self._conv2d_op = functools.partial(
tf.keras.layers.SeparableConv2D, depth_multiplier=1)
else:
self._conv2d_op = tf.keras.layers.Conv2D
if activation == 'relu':
self._activation_op = tf.nn.relu
elif activation == 'swish':
self._activation_op = tf.nn.swish
else:
raise ValueError('Unsupported activation `{}`.'.format(activation))
self._use_batch_norm = use_batch_norm
self._norm_activation = norm_activation
self._norm_activations = {}
self._lateral_conv2d_op = {}
self._post_hoc_conv2d_op = {}
self._coarse_conv2d_op = {}
for level in range(self._min_level, self._max_level + 1):
if self._use_batch_norm:
self._norm_activations[level] = norm_activation(
use_activation=False, name='p%d-bn' % level)
self._lateral_conv2d_op[level] = self._conv2d_op(
filters=self._fpn_feat_dims,
kernel_size=(1, 1),
padding='same',
name='l%d' % level)
self._post_hoc_conv2d_op[level] = self._conv2d_op(
filters=self._fpn_feat_dims,
strides=(1, 1),
kernel_size=(3, 3),
padding='same',
name='post_hoc_d%d' % level)
self._coarse_conv2d_op[level] = self._conv2d_op(
filters=self._fpn_feat_dims,
strides=(2, 2),
kernel_size=(3, 3),
padding='same',
name='p%d' % level)
