def _yolo_block(self, inputs, filters):
inputs = common._conv2d_fixed_padding(inputs, filters * 1, 1)
inputs = common._conv2d_fixed_padding(inputs, filters * 2, 3)
inputs = common._conv2d_fixed_padding(inputs, filters * 1, 1)
inputs = common._conv2d_fixed_padding(inputs, filters * 2, 3)
inputs = slim.conv2d(inputs, filters, 1, stride=1, padding='SAME', normalizer_fn=slim.batch_norm,
normalizer_params=self.batch_norm_params,
biases_initializer=None,
activation_fn=None)
return inputs
def _darknet53_block(self, inputs, filters):
"""
implement residuals block in darknet53
"""
shortcut = inputs
inputs = common._conv2d_fixed_padding(inputs, filters * 1, 1)
inputs = common._conv2d_fixed_padding(inputs, filters * 2, 3)
inputs = inputs + shortcut
return inputs
def forward(self, inputs, is_training=False, reuse=False):
"""
Creates YOLO v3 model.
:param inputs: a 4-D tensor of size [batch_size, height, width, channels].
Dimension batch_size may be undefined. The channel order is RGB.
:param is_training: whether is training or not.
:param reuse: whether or not the network and its variables should be reused.
:return:
"""
# it will be needed later on
self.img_size = tf.shape(inputs)[1:3]
# set batch norm params
batch_norm_params = {
'decay': self._BATCH_NORM_DECAY,
'epsilon': 1e-05,
'scale': True,
'is_training': is_training,
'fused': None, # Use fused batch norm if possible.
}
# Set activation_fn and parameters for conv2d, batch_norm.
with slim.arg_scope([slim.conv2d, slim.batch_norm, common._fixed_padding],reuse=reuse):
with slim.arg_scope([slim.conv2d], normalizer_fn=slim.batch_norm,
normalizer_params=batch_norm_params,
biases_initializer=None,
activation_fn=lambda x: tf.nn.leaky_relu(x, alpha=self._LEAKY_RELU)):
with tf.variable_scope('darknet-53'):
route_1, route_2, inputs = darknet53(inputs).outputs
with tf.variable_scope('yolo-v3'):
route, inputs = self._yolo_block(inputs, 512)
feature_map_1 = self._detection_layer(inputs, self._ANCHORS[6:9])
feature_map_1 = tf.identity(feature_map_1, name='feature_map_1')
inputs = common._conv2d_fixed_padding(route, 256, 1)
upsample_size = route_2.get_shape().as_list()
inputs = self._upsample(inputs, upsample_size)
inputs = tf.concat([inputs, route_2], axis=3)
route, inputs = self._yolo_block(inputs, 256)
feature_map_2 = self._detection_layer(inputs, self._ANCHORS[3:6])
feature_map_2 = tf.identity(feature_map_2, name='feature_map_2')
inputs = common._conv2d_fixed_padding(route, 128, 1)
upsample_size = route_1.get_shape().as_list()
inputs = self._upsample(inputs, upsample_size)
inputs = tf.concat([inputs, route_1], axis=3)
route, inputs = self._yolo_block(inputs, 128)
feature_map_3 = self._detection_layer(inputs, self._ANCHORS[0:3])
feature_map_3 = tf.identity(feature_map_3, name='feature_map_3')
return feature_map_1, feature_map_2, feature_map_3
def _darknet53_block(self, inputs, filters):
"""
#todo:darknet53_block残差模块
implement residuals block in darknet53
:param inputs: 输入数据
:param filters: 卷积核个数
:return: 两次卷积后数据变为尺寸变为原来的1/2，且是和输入数据有关的
"""
shortcut = inputs
inputs = common._conv2d_fixed_padding(inputs, filters * 1, 1) #same
inputs = common._conv2d_fixed_padding(inputs, filters * 2, 3) #下采样
inputs = inputs + shortcut
return inputs
def forward(self, inputs, n_filters, n_strides, n_ksizes):
for i in range(min(len(n_filters), len(n_strides)), len(n_ksizes)):
inputs = common._conv2d_fixed_padding(inputs,
filters=n_filters[i],
kernel_size=n_ksizes[i],
strides=n_strides[i])
return inputs
def forward(self, inputs):
inputs = common._conv2d_fixed_padding(inputs, 32, 3, strides=1)
inputs = common._conv2d_fixed_padding(inputs, 64, 3, strides=2)
inputs = self._darknet53_block(inputs, 32)
inputs = common._conv2d_fixed_padding(inputs, 128, 3, strides=2)
for i in range(2):
inputs = self._darknet53_block(inputs, 64)
inputs = common._conv2d_fixed_padding(inputs, 256, 3, strides=2)
for i in range(8):
inputs = self._darknet53_block(inputs, 128)
route_1 = inputs
inputs = common._conv2d_fixed_padding(inputs, 512, 3, strides=2)
for i in range(8):
inputs = self._darknet53_block(inputs, 256)
route_2 = inputs
inputs = common._conv2d_fixed_padding(inputs, 1024, 3, strides=2)
for i in range(4):
inputs = self._darknet53_block(inputs, 512)
return route_1, route_2, inputs
def _yolo_block(self, inputs, filters):
inputs = common._conv2d_fixed_padding(inputs, filters * 1, 1)
inputs = common._conv2d_fixed_padding(inputs, filters * 2, 3)
inputs = common._conv2d_fixed_padding(inputs, filters * 1, 1)
inputs = common._conv2d_fixed_padding(inputs, filters * 2, 3)
inputs = common._conv2d_fixed_padding(inputs, filters * 1, 1)
inputs = common._conv2d_fixed_padding(inputs, filters * 2, 3)
return inputs
def _yolo_block(self, inputs, filters):
# if stride > 1 , padding
inputs = common._conv2d_fixed_padding(inputs, filters * 1, 1)
inputs = common._conv2d_fixed_padding(inputs, filters * 2, 3)
inputs = common._conv2d_fixed_padding(inputs, filters * 1, 1)
inputs = common._conv2d_fixed_padding(inputs, filters * 2, 3)
inputs = common._conv2d_fixed_padding(inputs, filters * 1, 1)
route = inputs
inputs = common._conv2d_fixed_padding(inputs, filters * 2, 3)
return route, inputs
def _yolo_block(self, inputs, filters):
"""
yolo残差模块
:param inputs: 输入数据
:param filters: #卷积核个数
:return: route, inputs
"""
inputs = common._conv2d_fixed_padding(inputs, filters * 1, 1)
inputs = common._conv2d_fixed_padding(inputs, filters * 2, 3)
inputs = common._conv2d_fixed_padding(inputs, filters * 1, 1)
inputs = common._conv2d_fixed_padding(inputs, filters * 2, 3)
inputs = common._conv2d_fixed_padding(inputs, filters * 1, 1)
route = inputs
inputs = common._conv2d_fixed_padding(inputs, filters * 2, 3)
return route, inputs
def forward(self, inputs, is_training=False, reuse=False):
"""
Creates YOLO v3 model.
:param inputs: a 4-D tensor of size [batch_size, height, width, channels].
Dimension batch_size may be undefined. The channel order is RGB.
:param is_training: whether is training or not.
:param reuse: whether or not the network and its variables should be reused.
:return:
"""
# it will be needed later on 他在稍后将被需要
self.img_size = tf.shape(inputs)[1:3]
# set batch norm params
batch_norm_params = {
'decay': self.
_BATCH_NORM_DECAY, # https://www.cnblogs.com/hellcat/p/8058092.html
'epsilon': 1e-05,
'scale': True,
'is_training': is_training,
'fused': None, # Use fused batch norm if possible.
}
# Set activation_fn and parameters for conv2d, batch_norm.
with slim.arg_scope(
[slim.conv2d, slim.batch_norm, common._fixed_padding],
reuse=reuse):
with slim.arg_scope(
[slim.conv2d],
normalizer_fn=slim.batch_norm,
# 给定list(slim.conv2d)中的值设置默认值(normlizer,biase.....)
normalizer_params=batch_norm_params,
biases_initializer=None,
activation_fn=lambda x: tf.nn.leaky_relu(
x, alpha=self._LEAKY_RELU)):
with tf.variable_scope('darknet-53'):
route_1, route_2, inputs = darknet53(
inputs).outputs # 得到图片张量
# route_1 : 52x52x256
# route_2 : 26x26x512
# inputs : 13x13x1024
with tf.variable_scope('yolo-v3'):
# https://github.com/YunYang1994/tensorflow-yolov3/raw/master/docs/images/levio.jpeg
# https://images2018.cnblogs.com/blog/606386/201803/606386-20180327004340505-1572852891.png
# feature_map1 13x13x1024 --> 13x13x[3x(5+class_num)]
route, inputs = self._yolo_block(inputs, 512)
feature_map_1 = self._detection_layer(
inputs, self._ANCHORS[6:9])
feature_map_1 = tf.identity(feature_map_1,
name='feature_map_1')
# feature_map2 26x26x512 --> 26x26x[3x(5+class_num)]
inputs = common._conv2d_fixed_padding(route, 256, 1)
upsample_size = route_2.get_shape().as_list()
# 52x52 --> 26x26
inputs = self._upsample(inputs,
upsample_size) # 通过直接放大进行上采样
inputs = tf.concat([inputs, route_2],
axis=3) # 在axis=3 进行连接,
route, inputs = self._yolo_block(inputs, 256)
feature_map_2 = self._detection_layer(
inputs, self._ANCHORS[3:6])
feature_map_2 = tf.identity(feature_map_2,
name='feature_map_2')
# feature_map3 52x52x256 --> 52x52x[3x(5+class_num)]
inputs = common._conv2d_fixed_padding(route, 128, 1)
upsample_size = route_1.get_shape().as_list()
# 26x26 --> 52x52
inputs = self._upsample(inputs, upsample_size)
inputs = tf.concat([inputs, route_1], axis=3)
route, inputs = self._yolo_block(inputs, 128)
feature_map_3 = self._detection_layer(
inputs, self._ANCHORS[0:3])
feature_map_3 = tf.identity(feature_map_3,
name='feature_map_3')
return feature_map_1, feature_map_2, feature_map_3
