def test_get_boxes_for_five_aspect_ratios_per_location_fully_convolutional(
self):
image_features = tf.placeholder(dtype=tf.float32, shape=[4, None, None, 64])
conv_box_predictor = box_predictor.ConvolutionalBoxPredictor(
is_training=False,
num_classes=0,
conv_hyperparams=self._build_arg_scope_with_conv_hyperparams(),
min_depth=0,
max_depth=32,
num_layers_before_predictor=1,
use_dropout=True,
dropout_keep_prob=0.8,
kernel_size=1,
box_code_size=4
)
box_predictions = conv_box_predictor.predict(
image_features, num_predictions_per_location=5, scope='BoxPredictor')
box_encodings = box_predictions[box_predictor.BOX_ENCODINGS]
objectness_predictions = box_predictions[
box_predictor.CLASS_PREDICTIONS_WITH_BACKGROUND]
init_op = tf.global_variables_initializer()
resolution = 32
expected_num_anchors = resolution*resolution*5
with self.test_session() as sess:
sess.run(init_op)
(box_encodings_shape,
objectness_predictions_shape) = sess.run(
[tf.shape(box_encodings), tf.shape(objectness_predictions)],
feed_dict={image_features:
np.random.rand(4, resolution, resolution, 64)})
self.assertAllEqual(box_encodings_shape, [4, expected_num_anchors, 1, 4])
self.assertAllEqual(objectness_predictions_shape,
[4, expected_num_anchors, 1])
def test_get_multi_class_predictions_for_five_aspect_ratios_per_location(
self):
num_classes_without_background = 6
image_features = tf.random_uniform([4, 8, 8, 64], dtype=tf.float32)
conv_box_predictor = box_predictor.ConvolutionalBoxPredictor(
is_training=False,
num_classes=num_classes_without_background,
conv_hyperparams=self._build_arg_scope_with_conv_hyperparams(),
min_depth=0,
max_depth=32,
num_layers_before_predictor=1,
use_dropout=True,
dropout_keep_prob=0.8,
kernel_size=1,
box_code_size=4
)
box_predictions = conv_box_predictor.predict(
image_features,
num_predictions_per_location=5,
scope='BoxPredictor')
box_encodings = box_predictions[box_predictor.BOX_ENCODINGS]
class_predictions_with_background = box_predictions[
box_predictor.CLASS_PREDICTIONS_WITH_BACKGROUND]
init_op = tf.global_variables_initializer()
with self.test_session() as sess:
sess.run(init_op)
(box_encodings_shape, class_predictions_with_background_shape
) = sess.run([
tf.shape(box_encodings), tf.shape(class_predictions_with_background)])
self.assertAllEqual(box_encodings_shape, [4, 320, 1, 4])
self.assertAllEqual(class_predictions_with_background_shape,
[4, 320, num_classes_without_background+1])
def test_use_depthwise_convolution(self):
image_features = tf.placeholder(dtype=tf.float32,
shape=[4, None, None, 64])
conv_box_predictor = box_predictor.ConvolutionalBoxPredictor(
is_training=False,
num_classes=0,
conv_hyperparams_fn=self._build_arg_scope_with_conv_hyperparams(),
min_depth=0,
max_depth=32,
num_layers_before_predictor=1,
dropout_keep_prob=0.8,
kernel_size=1,
box_code_size=4,
use_dropout=True,
use_depthwise=True)
box_predictions = conv_box_predictor.predict(
[image_features],
num_predictions_per_location=[5],
scope='BoxPredictor')
box_encodings = tf.concat(box_predictions[box_predictor.BOX_ENCODINGS],
axis=1)
objectness_predictions = tf.concat(
box_predictions[box_predictor.CLASS_PREDICTIONS_WITH_BACKGROUND],
axis=1)
init_op = tf.global_variables_initializer()
resolution = 32
expected_num_anchors = resolution * resolution * 5
with self.test_session() as sess:
sess.run(init_op)
(box_encodings_shape, objectness_predictions_shape) = sess.run(
[tf.shape(box_encodings),
tf.shape(objectness_predictions)],
feed_dict={
image_features: np.random.rand(4, resolution, resolution,
64)
})
actual_variable_set = set(
[var.op.name for var in tf.trainable_variables()])
self.assertAllEqual(box_encodings_shape,
[4, expected_num_anchors, 1, 4])
self.assertAllEqual(objectness_predictions_shape,
[4, expected_num_anchors, 1])
expected_variable_set = set([
'BoxPredictor/Conv2d_0_1x1_32/biases',
'BoxPredictor/Conv2d_0_1x1_32/weights',
'BoxPredictor/BoxEncodingPredictor_depthwise/biases',
'BoxPredictor/BoxEncodingPredictor_depthwise/depthwise_weights',
'BoxPredictor/BoxEncodingPredictor/biases',
'BoxPredictor/BoxEncodingPredictor/weights',
'BoxPredictor/ClassPredictor_depthwise/biases',
'BoxPredictor/ClassPredictor_depthwise/depthwise_weights',
'BoxPredictor/ClassPredictor/biases',
'BoxPredictor/ClassPredictor/weights'
])
self.assertEqual(expected_variable_set, actual_variable_set)
def graph_fn(image_features):
conv_box_predictor = box_predictor.ConvolutionalBoxPredictor(
is_training=False,
num_classes=num_classes_without_background,
conv_hyperparams=self._build_arg_scope_with_conv_hyperparams(),
min_depth=0,
max_depth=32,
num_layers_before_predictor=1,
use_dropout=True,
dropout_keep_prob=0.8,
kernel_size=1,
box_code_size=4)
box_predictions = conv_box_predictor.predict(
[image_features],
num_predictions_per_location=[5],
scope='BoxPredictor')
box_encodings = box_predictions[box_predictor.BOX_ENCODINGS]
class_predictions_with_background = box_predictions[
box_predictor.CLASS_PREDICTIONS_WITH_BACKGROUND]
return (box_encodings, class_predictions_with_background)
def build(argscope_fn, box_predictor_config, is_training, num_classes):
"""Builds box predictor based on the configuration.
Builds box predictor based on the configuration. See box_predictor.proto for
configurable options. Also, see box_predictor.py for more details.
Args:
argscope_fn: A function that takes the following inputs:
* hyperparams_pb2.Hyperparams proto
* a boolean indicating if the model is in training mode.
and returns a tf slim argscope for Conv and FC hyperparameters.
box_predictor_config: box_predictor_pb2.BoxPredictor proto containing
configuration.
is_training: Whether the models is in training mode.
num_classes: Number of classes to predict.
Returns:
box_predictor: box_predictor.BoxPredictor object.
Raises:
ValueError: On unknown box predictor.
"""
if not isinstance(box_predictor_config, box_predictor_pb2.BoxPredictor):
raise ValueError('box_predictor_config not of type '
'box_predictor_pb2.BoxPredictor.')
box_predictor_oneof = box_predictor_config.WhichOneof('box_predictor_oneof')
if box_predictor_oneof == 'convolutional_box_predictor':
conv_box_predictor = box_predictor_config.convolutional_box_predictor
conv_hyperparams = argscope_fn(conv_box_predictor.conv_hyperparams,
is_training)
box_predictor_object = box_predictor.ConvolutionalBoxPredictor(
is_training=is_training,
num_classes=num_classes,
conv_hyperparams=conv_hyperparams,
min_depth=conv_box_predictor.min_depth,
max_depth=conv_box_predictor.max_depth,
num_layers_before_predictor=(conv_box_predictor.
num_layers_before_predictor),
use_dropout=conv_box_predictor.use_dropout,
dropout_keep_prob=conv_box_predictor.dropout_keep_probability,
kernel_size=conv_box_predictor.kernel_size,
box_code_size=conv_box_predictor.box_code_size,
apply_sigmoid_to_scores=conv_box_predictor.apply_sigmoid_to_scores)
return box_predictor_object
if box_predictor_oneof == 'mask_rcnn_box_predictor':
mask_rcnn_box_predictor = box_predictor_config.mask_rcnn_box_predictor
fc_hyperparams = argscope_fn(mask_rcnn_box_predictor.fc_hyperparams,
is_training)
conv_hyperparams = None
if mask_rcnn_box_predictor.HasField('conv_hyperparams'):
conv_hyperparams = argscope_fn(mask_rcnn_box_predictor.conv_hyperparams,
is_training)
box_predictor_object = box_predictor.MaskRCNNBoxPredictor(
is_training=is_training,
num_classes=num_classes,
fc_hyperparams=fc_hyperparams,
use_dropout=mask_rcnn_box_predictor.use_dropout,
dropout_keep_prob=mask_rcnn_box_predictor.dropout_keep_probability,
box_code_size=mask_rcnn_box_predictor.box_code_size,
conv_hyperparams=conv_hyperparams,
predict_instance_masks=mask_rcnn_box_predictor.predict_instance_masks,
mask_prediction_conv_depth=(mask_rcnn_box_predictor.
mask_prediction_conv_depth),
predict_keypoints=mask_rcnn_box_predictor.predict_keypoints)
return box_predictor_object
if box_predictor_oneof == 'rfcn_box_predictor':
rfcn_box_predictor = box_predictor_config.rfcn_box_predictor
conv_hyperparams = argscope_fn(rfcn_box_predictor.conv_hyperparams,
is_training)
box_predictor_object = box_predictor.RfcnBoxPredictor(
is_training=is_training,
num_classes=num_classes,
conv_hyperparams=conv_hyperparams,
crop_size=[rfcn_box_predictor.crop_height,
rfcn_box_predictor.crop_width],
num_spatial_bins=[rfcn_box_predictor.num_spatial_bins_height,
rfcn_box_predictor.num_spatial_bins_width],
depth=rfcn_box_predictor.depth,
box_code_size=rfcn_box_predictor.box_code_size)
return box_predictor_object
raise ValueError('Unknown box predictor: {}'.format(box_predictor_oneof))
def build(argscope_fn, box_predictor_config, is_training, num_classes):
"""Builds box predictor based on the configuration.
Builds box predictor based on the configuration. See box_predictor.proto for
configurable options. Also, see box_predictor.py for more details.
Args:
argscope_fn: A function that takes the following inputs:
* hyperparams_pb2.Hyperparams proto
* a boolean indicating if the model is in training mode.
and returns a tf slim argscope for Conv and FC hyperparameters.
box_predictor_config: box_predictor_pb2.BoxPredictor proto containing
configuration.
is_training: Whether the models is in training mode.
num_classes: Number of classes to predict.
Returns:
box_predictor: box_predictor.BoxPredictor object.
Raises:
ValueError: On unknown box predictor.
"""
if not isinstance(box_predictor_config, box_predictor_pb2.BoxPredictor):
raise ValueError('box_predictor_config not of type '
'box_predictor_pb2.BoxPredictor.')
box_predictor_oneof = box_predictor_config.WhichOneof(
'box_predictor_oneof') #not giben in the parameter
if box_predictor_oneof == 'convolutional_box_predictor': #we don't have this(faster RCNN) . I think this used for SSD
conv_box_predictor = box_predictor_config.convolutional_box_predictor #get the params
conv_hyperparams = argscope_fn(
conv_box_predictor.
conv_hyperparams, #converting the convolution hyper params needed for the additional conv
is_training
) #convolution and making help to make the arg scope . done with hyperparametr_builder.build
#This is what following object can do : Take high level image feature map as input and produce two predictions,
#(1) a tensor encoding box locations, and
#(2) a tensor encoding classes for each box. - from the box_predictor.py script
#These components are passed directly to loss functions
box_predictor_object = box_predictor.ConvolutionalBoxPredictor( #Initialize the boxpredictor class
is_training=is_training,
num_classes=num_classes,
conv_hyperparams=
conv_hyperparams, #above slim arg scope which have weight , operations etc
min_depth=conv_box_predictor.min_depth,
max_depth=conv_box_predictor.max_depth,
num_layers_before_predictor=(
conv_box_predictor. #how many layers before the predictoot
num_layers_before_predictor),
use_dropout=conv_box_predictor.use_dropout,
dropout_keep_prob=conv_box_predictor.dropout_keep_probability,
kernel_size=conv_box_predictor.kernel_size,
box_code_size=conv_box_predictor.
box_code_size, #what is this size of the encoding ??????????????
apply_sigmoid_to_scores=conv_box_predictor.apply_sigmoid_to_scores)
return box_predictor_object #retunrs this since box _predictor class has many functiosn
if box_predictor_oneof == 'mask_rcnn_box_predictor': #this is what we have for faster Rcnn
mask_rcnn_box_predictor = box_predictor_config.mask_rcnn_box_predictor #get the hyperparams
fc_hyperparams = argscope_fn(
mask_rcnn_box_predictor.
fc_hyperparams, #this is basically getting fully connected params in order to predict
is_training)
conv_hyperparams = None
if mask_rcnn_box_predictor.HasField(
'conv_hyperparams'): #don't have one yet
conv_hyperparams = argscope_fn(
mask_rcnn_box_predictor.
conv_hyperparams, #assigning hyper params
is_training)
box_predictor_object = box_predictor.MaskRCNNBoxPredictor( #this create a object from box_predictor class which use to initialize thing ss
is_training=is_training,
num_classes=num_classes,
fc_hyperparams=fc_hyperparams,
use_dropout=mask_rcnn_box_predictor.use_dropout,
dropout_keep_prob=mask_rcnn_box_predictor.dropout_keep_probability,
box_code_size=mask_rcnn_box_predictor.box_code_size,
conv_hyperparams=conv_hyperparams,
predict_instance_masks=mask_rcnn_box_predictor.
predict_instance_masks,
mask_prediction_conv_depth=(
mask_rcnn_box_predictor.mask_prediction_conv_depth),
predict_keypoints=mask_rcnn_box_predictor.predict_keypoints)
return box_predictor_object
if box_predictor_oneof == 'rfcn_box_predictor': #if rfcn we get thus
rfcn_box_predictor = box_predictor_config.rfcn_box_predictor
conv_hyperparams = argscope_fn(
rfcn_box_predictor.
conv_hyperparams, #then we get this slim.arguments
is_training)
box_predictor_object = box_predictor.RfcnBoxPredictor( #again this is for rfcn same as above
is_training=is_training,
num_classes=num_classes,
conv_hyperparams=conv_hyperparams,
crop_size=[
rfcn_box_predictor.crop_height, rfcn_box_predictor.crop_width
],
num_spatial_bins=[
rfcn_box_predictor.num_spatial_bins_height,
rfcn_box_predictor.num_spatial_bins_width
],
depth=rfcn_box_predictor.depth,
box_code_size=rfcn_box_predictor.box_code_size)
return box_predictor_object
raise ValueError('Unknown box predictor: {}'.format(box_predictor_oneof))
