def _instantiate_layers(self):
"""
定义两个全连接层和后面的分类回归层
:return: 无
"""
# We define layers as an array since they are simple fully connected
# ones and it should be easy to tune it from the network config.
# 定义两个全连接层, layer_size表示输出通道数
self._layers = [
snt.Linear(
layer_size,
name='fc_{}'.format(i),
initializers={'w': self._rcnn_initializer},
regularizers={'w': self.regularizer},
) for i, layer_size in enumerate(self._layer_sizes)
]
# We define the classifier layer having a num_classes + 1 background
# since we want to be able to predict if the proposal is background as
# well.
# 定义一个分类的全连接层(21类)
self._classifier_layer = snt.Linear(
self._num_classes + 1,
name='fc_classifier',
initializers={'w': self._cls_initializer},
regularizers={'w': self.regularizer},
)
# The bounding box adjustment layer has 4 times the number of classes
# We choose which to use depending on the output of the classifier
# layer
# 输出通道数为类别数目x4, 也就是4x20, 每个类别对应4个值, 这个是独立于分类器的输出
self._bbox_layer = snt.Linear(
self._num_classes * 4,
name='fc_bbox',
initializers={'w': self._bbox_initializer},
regularizers={'w': self.regularizer})
# ROIPoolingLayer is used to extract the feature from the feature map
# using the proposals.
# self._config.roi表示的就是选择crop, 7x7, valid的池化配置
self._roi_pool = ROIPoolingLayer(self._config.roi, debug=self._debug)
# RCNNTarget is used to define a minibatch and the correct values for
# each of the proposals.
# 用来定义minibatch和提案的正确值
self._rcnn_target = RCNNTarget(self._num_classes,
self._config.target,
variances=self._variances,
seed=self._seed)
# RCNNProposal generates the final bounding boxes and tries to remove
# duplicates.
self._rcnn_proposal = RCNNProposal(self._num_classes,
self._config.proposals,
variances=self._variances)
def testLimits(self):
"""Tests that we're respecting the limits imposed by the config.
"""
limits_config = self._config.copy()
limits_config['class_max_detections'] = 2
limits_config['total_max_detections'] = 3
limits_config = EasyDict(limits_config)
limits_num_classes = 2
limits_model = RCNNProposal(limits_num_classes, limits_config)
proposed_boxes = tf.constant([
(self._batch_number, 0, 0, 1, 1), # class 0
(self._batch_number, 5, 5, 10, 10), # class 1
(self._batch_number, 15, 15, 20, 20), # class 1
(self._batch_number, 25, 25, 30, 30), # class 0
(self._batch_number, 35, 35, 40, 40),
(self._batch_number, 38, 40, 65, 65),
(self._batch_number, 70, 50, 90, 90), # class 0
(self._batch_number, 95, 95, 100, 100),
(self._batch_number, 105, 105, 110, 110), # class 1
])
# All zeroes for our bbox_pred.
bbox_pred = tf.constant([[0.] * limits_num_classes * 4] * 9)
cls_prob = tf.constant([
(0., 1., 0.),
(0., .2, .8),
(0., .45, .55),
(0., .55, .45),
(1., 0., 0.),
(1., 0., 0.),
(0., .95, .05),
(1., 0., 0.),
(0., .495, .505),
])
proposal_prediction = self._run_rcnn_proposal(
limits_model,
proposed_boxes,
bbox_pred,
cls_prob,
)
labels = proposal_prediction['proposal_label']
num_class0 = labels[labels == 0].shape[0]
num_class1 = labels[labels == 1].shape[0]
self.assertLessEqual(num_class0, limits_config.class_max_detections)
self.assertLessEqual(num_class1, limits_config.class_max_detections)
num_total = labels.shape[0]
self.assertLessEqual(num_total, limits_config.total_max_detections)
def _instantiate_layers(self):
# We define layers as an array since they are simple fully connected
# ones and it should be easy to tune it from the network config.
self._layers = [
snt.Linear(
layer_size,
name="fc_{}".format(i),
initializers={"w": self._rcnn_initializer},
regularizers={"w": self.regularizer},
)
for i, layer_size in enumerate(self._layer_sizes)
]
# We define the classifier layer having a num_classes + 1 background
# since we want to be able to predict if the proposal is background as
# well.
self._classifier_layer = snt.Linear(
self._num_classes + 1,
name="fc_classifier",
initializers={"w": self._cls_initializer},
regularizers={"w": self.regularizer},
)
# The bounding box adjustment layer has 4 times the number of classes
# We choose which to use depending on the output of the classifier
# layer
self._bbox_layer = snt.Linear(
self._num_classes * 4,
name="fc_bbox",
initializers={"w": self._bbox_initializer},
regularizers={"w": self.regularizer},
)
# ROIPoolingLayer is used to extract the feature from the feature map
# using the proposals.
self._roi_pool = ROIPoolingLayer(self._config.roi, debug=self._debug)
# RCNNTarget is used to define a minibatch and the correct values for
# each of the proposals.
self._rcnn_target = RCNNTarget(
self._num_classes,
self._config.target,
variances=self._variances,
seed=self._seed,
)
# RCNNProposal generates the final bounding boxes and tries to remove
# duplicates.
self._rcnn_proposal = RCNNProposal(
self._num_classes, self._config.proposals, variances=self._variances
)
def setUp(self):
super(RCNNProposalTest, self).setUp()
self._num_classes = 3
self._image_shape = (900, 1440)
self._config = EasyDict({
'class_max_detections': 100,
'class_nms_threshold': 0.6,
'total_max_detections': 300,
'min_prob_threshold': 0.0,
})
self._equality_delta = 1e-03
self._shared_model = RCNNProposal(self._num_classes, self._config)
tf.reset_default_graph()