def __build_nework(self, input_data):
"""
It is better look at this function with the Yolo_v3_Structure diagram made by Levio
One link is here: https://github.com/wizyoung/YOLOv3_TensorFlow
"""
# by YZ, these are the three feature extractions from darknet53, with difference dimension reductions (by 8, 16, 32)
route_1, route_2, input_data = backbone.darknet53(input_data, self.trainable)
# input_data is -1*13*13*1024
input_data = common.convolutional(input_data, (1, 1, 1024, 512), self.trainable, 'conv52')
input_data = common.convolutional(input_data, (3, 3, 512, 1024), self.trainable, 'conv53')
input_data = common.convolutional(input_data, (1, 1, 1024, 512), self.trainable, 'conv54')
input_data = common.convolutional(input_data, (3, 3, 512, 1024), self.trainable, 'conv55')
input_data = common.convolutional(input_data, (1, 1, 1024, 512), self.trainable, 'conv56')
conv_lobj_branch = common.convolutional(input_data, (3, 3, 512, 1024), self.trainable, name='conv_lobj_branch')
conv_lbbox = common.convolutional(conv_lobj_branch, (1, 1, 1024, 3*(self.num_class + 5)),
trainable=self.trainable, name='conv_lbbox', activate=False, bn=False)
input_data = common.convolutional(input_data, (1, 1, 512, 256), self.trainable, 'conv57')
# upsampling input data (1/32) to match route_2 (1/16), -1*26*16*512
input_data = common.upsample(input_data, name='upsample0', method=self.upsample_method)
with tf.variable_scope('route_1'):
# route_2 is -1*26*26*256, so input_data is -1*26*26*768
input_data = tf.concat([input_data, route_2], axis=-1)
input_data = common.convolutional(input_data, (1, 1, 768, 256), self.trainable, 'conv58')
input_data = common.convolutional(input_data, (3, 3, 256, 512), self.trainable, 'conv59')
input_data = common.convolutional(input_data, (1, 1, 512, 256), self.trainable, 'conv60')
input_data = common.convolutional(input_data, (3, 3, 256, 512), self.trainable, 'conv61')
input_data = common.convolutional(input_data, (1, 1, 512, 256), self.trainable, 'conv62')
conv_mobj_branch = common.convolutional(input_data, (3, 3, 256, 512), self.trainable, name='conv_mobj_branch' )
conv_mbbox = common.convolutional(conv_mobj_branch, (1, 1, 512, 3*(self.num_class + 5)),
trainable=self.trainable, name='conv_mbbox', activate=False, bn=False)
input_data = common.convolutional(input_data, (1, 1, 256, 128), self.trainable, 'conv63')
input_data = common.upsample(input_data, name='upsample1', method=self.upsample_method)
with tf.variable_scope('route_2'):
input_data = tf.concat([input_data, route_1], axis=-1)
input_data = common.convolutional(input_data, (1, 1, 384, 128), self.trainable, 'conv64')
input_data = common.convolutional(input_data, (3, 3, 128, 256), self.trainable, 'conv65')
input_data = common.convolutional(input_data, (1, 1, 256, 128), self.trainable, 'conv66')
input_data = common.convolutional(input_data, (3, 3, 128, 256), self.trainable, 'conv67')
input_data = common.convolutional(input_data, (1, 1, 256, 128), self.trainable, 'conv68')
conv_sobj_branch = common.convolutional(input_data, (3, 3, 128, 256), self.trainable, name='conv_sobj_branch')
conv_sbbox = common.convolutional(conv_sobj_branch, (1, 1, 256, 3*(self.num_class + 5)),
trainable=self.trainable, name='conv_sbbox', activate=False, bn=False)
# dimensions are: -1*13*13*255, -1*26*26*255, -1*52*52*255
return conv_lbbox, conv_mbbox, conv_sbbox
def __build_nework(self, input_layer, training):
route_1, route_2, conv = backbone.darknet53(input_layer, training)
conv = common.convolutional(conv, (1, 1, 1024, 512), training)
conv = common.convolutional(conv, (3, 3, 512, 1024), training)
conv = common.convolutional(conv, (1, 1, 1024, 512), training)
conv = common.convolutional(conv, (3, 3, 512, 1024), training)
conv = common.convolutional(conv, (1, 1, 1024, 512), training)
conv_lobj_branch = common.convolutional(conv, (3, 3, 512, 1024),
training)
conv_lbbox = common.convolutional(conv_lobj_branch,
(1, 1, 1024, 3 * (80 + 5)),
training,
activate=False,
bn=False)
conv = common.convolutional(conv, (1, 1, 512, 256), training)
conv = common.upsample(conv)
conv = tf.concat([conv, route_2], axis=-1)
conv = common.convolutional(conv, (1, 1, 768, 256), training)
conv = common.convolutional(conv, (3, 3, 256, 512), training)
conv = common.convolutional(conv, (1, 1, 512, 256), training)
conv = common.convolutional(conv, (3, 3, 256, 512), training)
conv = common.convolutional(conv, (1, 1, 512, 256), training)
conv_mobj_branch = common.convolutional(conv, (3, 3, 256, 512),
training)
conv_mbbox = common.convolutional(conv_mobj_branch,
(1, 1, 512, 3 * (80 + 5)),
training,
activate=False,
bn=False)
conv = common.convolutional(conv, (1, 1, 256, 128), training)
conv = common.upsample(conv)
conv = tf.concat([conv, route_1], axis=-1)
conv = common.convolutional(conv, (1, 1, 384, 128), training)
conv = common.convolutional(conv, (3, 3, 128, 256), training)
conv = common.convolutional(conv, (1, 1, 256, 128), training)
conv = common.convolutional(conv, (3, 3, 128, 256), training)
conv = common.convolutional(conv, (1, 1, 256, 128), training)
conv_sobj_branch = common.convolutional(conv, (3, 3, 128, 256),
training)
conv_sbbox = common.convolutional(conv_sobj_branch,
(1, 1, 256, 3 * (80 + 5)),
training,
activate=False,
bn=False)
return conv_lbbox, conv_mbbox, conv_sbbox
def YOLOv4(input_layer, NUM_CLASS):
route_1, route_2, conv = backbone.cspdarknet53(input_layer)
route = conv
conv = common.convolutional(conv, (1, 1, 512, 256))
conv = common.upsample(conv)
route_2 = common.convolutional(route_2, (1, 1, 512, 256))
conv = tf.concat([route_2, conv], axis=-1)
conv = common.convolutional(conv, (1, 1, 512, 256))
conv = common.convolutional(conv, (3, 3, 256, 512))
conv = common.convolutional(conv, (1, 1, 512, 256))
conv = common.convolutional(conv, (3, 3, 256, 512))
conv = common.convolutional(conv, (1, 1, 512, 256))
route_2 = conv
conv = common.convolutional(conv, (1, 1, 256, 128))
conv = common.upsample(conv)
route_1 = common.convolutional(route_1, (1, 1, 256, 128))
conv = tf.concat([route_1, conv], axis=-1)
conv = common.convolutional(conv, (1, 1, 256, 128))
conv = common.convolutional(conv, (3, 3, 128, 256))
conv = common.convolutional(conv, (1, 1, 256, 128))
conv = common.convolutional(conv, (3, 3, 128, 256))
conv = common.convolutional(conv, (1, 1, 256, 128))
route_1 = conv
conv = common.convolutional(conv, (3, 3, 128, 256))
conv_sbbox = common.convolutional(conv, (1, 1, 256, 3 * (NUM_CLASS + 5)), activate=False, bn=False)
conv = common.convolutional(route_1, (3, 3, 128, 256), downsample=True)
conv = tf.concat([conv, route_2], axis=-1)
conv = common.convolutional(conv, (1, 1, 512, 256))
conv = common.convolutional(conv, (3, 3, 256, 512))
conv = common.convolutional(conv, (1, 1, 512, 256))
conv = common.convolutional(conv, (3, 3, 256, 512))
conv = common.convolutional(conv, (1, 1, 512, 256))
route_2 = conv
conv = common.convolutional(conv, (3, 3, 256, 512))
conv_mbbox = common.convolutional(conv, (1, 1, 512, 3 * (NUM_CLASS + 5)), activate=False, bn=False)
conv = common.convolutional(route_2, (3, 3, 256, 512), downsample=True)
conv = tf.concat([conv, route], axis=-1)
conv = common.convolutional(conv, (1, 1, 1024, 512))
conv = common.convolutional(conv, (3, 3, 512, 1024))
conv = common.convolutional(conv, (1, 1, 1024, 512))
conv = common.convolutional(conv, (3, 3, 512, 1024))
conv = common.convolutional(conv, (1, 1, 1024, 512))
conv = common.convolutional(conv, (3, 3, 512, 1024))
conv_lbbox = common.convolutional(conv, (1, 1, 1024, 3 * (NUM_CLASS + 5)), activate=False, bn=False)
return [conv_sbbox, conv_mbbox, conv_lbbox]
def YOLOv3(for_annother_use=False):
input_layer = tf.keras.layers.Input(shape=[416,416,3])
route_1, route_2, conv = backbone.darknet53(input_layer)
conv = common.convolutional(conv, (1, 1, 1024, 512))
sconv1 = conv = common.convolutional(conv, (3, 3, 512, 1024))
conv = common.convolutional(conv, (1, 1, 1024, 512))
sconv2 = conv = common.convolutional(conv, (3, 3, 512, 1024))
conv = common.convolutional(conv, (1, 1, 1024, 512))
sconv3 = conv_lobj_branch = common.convolutional(conv, (3, 3, 512, 1024))
conv_lbbox = common.convolutional(conv_lobj_branch, (1, 1, 1024, 3 * (NUM_CLASS + 5)), activate=False, bn=False)
conv = common.convolutional(conv, (1, 1, 512, 256))
conv = common.upsample(conv)
conv = tf.concat([conv, route_2], axis=-1)
conv = common.convolutional(conv, (1, 1, 768, 256))
mconv1 = conv = common.convolutional(conv, (3, 3, 256, 512))
conv = common.convolutional(conv, (1, 1, 512, 256))
mconv2 = conv = common.convolutional(conv, (3, 3, 256, 512))
conv = common.convolutional(conv, (1, 1, 512, 256))
mconv3 = conv_mobj_branch = common.convolutional(conv, (3, 3, 256, 512))
conv_mbbox = common.convolutional(conv_mobj_branch, (1, 1, 512, 3 * (NUM_CLASS + 5)), activate=False, bn=False)
conv = common.convolutional(conv, (1, 1, 256, 128))
conv = common.upsample(conv)
conv = tf.concat([conv, route_1], axis=-1)
conv = common.convolutional(conv, (1, 1, 384, 128))
lconv1 = conv = common.convolutional(conv, (3, 3, 128, 256))
conv = common.convolutional(conv, (1, 1, 256, 128))
lconv2 = conv = common.convolutional(conv, (3, 3, 128, 256))
conv = common.convolutional(conv, (1, 1, 256, 128))
lconv3 = conv_sobj_branch = common.convolutional(conv, (3, 3, 128, 256))
conv_sbbox = common.convolutional(conv_sobj_branch, (1, 1, 256, 3 * (NUM_CLASS + 5)), activate=False, bn=False)
if for_annother_use:
convs_sbbox = tf.concat([lconv1, lconv2, lconv3], axis=-1, name="lfeature")
convs_mbbox = tf.concat([mconv1, mconv2, mconv3], axis=-1, name="mfeature")
convs_lbbox = tf.concat([sconv1, sconv2, sconv3], axis=-1, name="sfeature")
yolo = Model(input_layer,[conv_sbbox, conv_mbbox, conv_lbbox, convs_sbbox,convs_mbbox,convs_lbbox])
# 52,26,13 52,26,13
yolo.load_weights(filepath=tf.keras.utils.get_file('yolo3.h5',''))
yolo.trainable = False
else:
# classic YOLOV3 model trained for coco datasets
yolo = Model(input_layer, [conv_sbbox, conv_mbbox, conv_lbbox])
yolo.load_weights(filepath=tf.keras.utils.get_file('yolo3.h5',''))
yolo.trainable =False
return yolo
def __build_nework(self, input_data):
route_1, route_2, input_data = backbone.darknet53(input_data, self.trainable)
# input_data is -1*13*13*1024
input_data = common.convolutional(input_data, (1, 1, 1024, 512), self.trainable, 'conv52')
input_data = common.convolutional(input_data, (3, 3, 512, 1024), self.trainable, 'conv53')
input_data = common.convolutional(input_data, (1, 1, 1024, 512), self.trainable, 'conv54')
input_data = common.convolutional(input_data, (3, 3, 512, 1024), self.trainable, 'conv55')
input_data = common.convolutional(input_data, (1, 1, 1024, 512), self.trainable, 'conv56')
conv_lobj_branch = common.convolutional(input_data, (3, 3, 512, 1024), self.trainable, name='conv_lobj_branch')
conv_lbbox = common.convolutional(conv_lobj_branch, (1, 1, 1024, 3*(self.num_class + 5)),
trainable=self.trainable, name='conv_lbbox', activate=False, bn=False)
input_data = common.convolutional(input_data, (1, 1, 512, 256), self.trainable, 'conv57')
# upsampling input data (1/32) to match route_2 (1/16), -1*26*16*512
input_data = common.upsample(input_data, name='upsample0', method=self.upsample_method)
with tf.variable_scope('route_1'):
# route_2 is -1*26*26*256, so input_data is -1*26*26*768
input_data = tf.concat([input_data, route_2], axis=-1)
input_data = common.convolutional(input_data, (1, 1, 768, 256), self.trainable, 'conv58')
input_data = common.convolutional(input_data, (3, 3, 256, 512), self.trainable, 'conv59')
input_data = common.convolutional(input_data, (1, 1, 512, 256), self.trainable, 'conv60')
input_data = common.convolutional(input_data, (3, 3, 256, 512), self.trainable, 'conv61')
input_data = common.convolutional(input_data, (1, 1, 512, 256), self.trainable, 'conv62')
conv_mobj_branch = common.convolutional(input_data, (3, 3, 256, 512), self.trainable, name='conv_mobj_branch' )
conv_mbbox = common.convolutional(conv_mobj_branch, (1, 1, 512, 3*(self.num_class + 5)),
trainable=self.trainable, name='conv_mbbox', activate=False, bn=False)
input_data = common.convolutional(input_data, (1, 1, 256, 128), self.trainable, 'conv63')
input_data = common.upsample(input_data, name='upsample1', method=self.upsample_method)
with tf.variable_scope('route_2'):
input_data = tf.concat([input_data, route_1], axis=-1)
input_data = common.convolutional(input_data, (1, 1, 384, 128), self.trainable, 'conv64')
input_data = common.convolutional(input_data, (3, 3, 128, 256), self.trainable, 'conv65')
input_data = common.convolutional(input_data, (1, 1, 256, 128), self.trainable, 'conv66')
input_data = common.convolutional(input_data, (3, 3, 128, 256), self.trainable, 'conv67')
input_data = common.convolutional(input_data, (1, 1, 256, 128), self.trainable, 'conv68')
conv_sobj_branch = common.convolutional(input_data, (3, 3, 128, 256), self.trainable, name='conv_sobj_branch')
conv_sbbox = common.convolutional(conv_sobj_branch, (1, 1, 256, 3*(self.num_class + 5)),
trainable=self.trainable, name='conv_sbbox', activate=False, bn=False)
# dimensions are: -1*13*13*255, -1*26*26*255, -1*52*52*255
return conv_lbbox, conv_mbbox, conv_sbbox
def YOLOv3(input_layer):
"""
build the whole network of YOLOv3
:param input_layer: [batch, 416, 416, 3]
:return: 3 scale predictions [[b, 52, 52, anchors*(5+classes)], [b, 26, 26, anchors*(5+classes)], [b, 13, 13, anchors*(5+classes)]]
"""
route_1, route_2, conv = backbone.darknet53(
input_layer
) # call darknet to get the basic network, [52,52,256], [26,26,512], [13,13,1024]
"""large scale output [batch, 13, 13, 3*(5+c)]"""
conv = common.convolutional(conv, (1, 1, 1024, 512), conv_trainable=False)
conv = common.convolutional(conv, (3, 3, 512, 1024), conv_trainable=False)
conv = common.convolutional(conv, (1, 1, 1024, 512), conv_trainable=False)
conv = common.convolutional(conv, (3, 3, 512, 1024))
conv = common.convolutional(conv, (1, 1, 1024, 512))
conv_lobj_branch = common.convolutional(conv, (3, 3, 512, 1024))
conv_lbbox = common.convolutional(conv_lobj_branch,
(1, 1, 1024, 3 * (NUM_CLASS + 5)),
activate=False,
bn=False)
"""the medium scale output, [batch, 26, 26, 3*(NUM_CLASS+5)]"""
conv = common.convolutional(conv, (1, 1, 512, 256), conv_trainable=False)
conv = common.upsample(conv)
conv = tf.concat([conv, route_2], axis=-1) # concat in channel dimension
conv = common.convolutional(conv, (1, 1, 768, 256), conv_trainable=False)
conv = common.convolutional(conv, (3, 3, 256, 512), conv_trainable=False)
conv = common.convolutional(conv, (1, 1, 512, 256), conv_trainable=False)
conv = common.convolutional(conv, (3, 3, 256, 512))
conv = common.convolutional(conv, (1, 1, 512, 256))
conv_mobj_branch = common.convolutional(conv, (3, 3, 256, 512))
conv_mbbox = common.convolutional(conv_mobj_branch,
(1, 1, 512, 3 * (NUM_CLASS + 5)),
activate=False,
bn=False)
"""the small scale output, [batch, 52, 52, 3*(NUM_CLASS+5)]"""
conv = common.convolutional(conv, (1, 1, 256, 128), conv_trainable=False)
conv = common.upsample(conv)
conv = tf.concat([conv, route_1], axis=-1) # concat in channel dimension
conv = common.convolutional(conv, (1, 1, 384, 128), conv_trainable=False)
conv = common.convolutional(conv, (3, 3, 128, 256), conv_trainable=False)
conv = common.convolutional(conv, (1, 1, 256, 128), conv_trainable=False)
conv = common.convolutional(conv, (3, 3, 128, 256))
conv = common.convolutional(conv, (1, 1, 256, 128))
conv_sobj_branch = common.convolutional(conv, (3, 3, 128, 256))
conv_sbbox = common.convolutional(conv_sobj_branch,
(1, 1, 256, 3 * (NUM_CLASS + 5)),
activate=False,
bn=False)
return [conv_sbbox, conv_mbbox, conv_lbbox]
def YOLOv3(input_layer):
route_1, route_2, conv = backbone.darknet53(input_layer)
original_conv = conv
conv = common.convolutional(conv, (1, 1, 1024, 512))
conv = common.convolutional(conv, (3, 3, 512, 1024))
conv = common.convolutional(conv, (1, 1, 1024, 512))
conv = common.convolutional(conv, (3, 3, 512, 1024))
conv = common.convolutional(conv, (1, 1, 1024, 512))
conv_lobj_branch = common.convolutional(conv, (3, 3, 512, 1024))
conv_lbbox = common.convolutional(conv_lobj_branch,
(1, 1, 1024, 3 * (NUM_CLASS + 5)),
activate=False,
bn=False)
conv = common.convolutional(conv, (1, 1, 512, 256))
conv = common.upsample(conv)
conv = tf.concat([conv, route_2], axis=-1)
conv = common.convolutional(conv, (1, 1, 768, 256))
conv = common.convolutional(conv, (3, 3, 256, 512))
conv = common.convolutional(conv, (1, 1, 512, 256))
conv = common.convolutional(conv, (3, 3, 256, 512))
conv = common.convolutional(conv, (1, 1, 512, 256))
conv_mobj_branch = common.convolutional(conv, (3, 3, 256, 512))
conv_mbbox = common.convolutional(conv_mobj_branch,
(1, 1, 512, 3 * (NUM_CLASS + 5)),
activate=False,
bn=False)
conv = common.convolutional(conv, (1, 1, 256, 128))
conv = common.upsample(conv)
conv = tf.concat([conv, route_1], axis=-1)
conv = common.convolutional(conv, (1, 1, 384, 128))
conv = common.convolutional(conv, (3, 3, 128, 256))
conv = common.convolutional(conv, (1, 1, 256, 128))
conv = common.convolutional(conv, (3, 3, 128, 256))
conv = common.convolutional(conv, (1, 1, 256, 128))
conv_sobj_branch = common.convolutional(conv, (3, 3, 128, 256))
conv_sbbox = common.convolutional(conv_sobj_branch,
(1, 1, 256, 3 * (NUM_CLASS + 5)),
activate=False,
bn=False)
return [conv_sbbox, conv_mbbox,
conv_lbbox], (route_1, route_2, original_conv)
def __build_nework(self, input_data):
route_1, route_2, input_data = backbone.darknet53(input_data)
input_data = common.convolutional(input_data, (1, 1, 1024, 512))
input_data = common.convolutional(input_data, (3, 3, 512, 1024))
input_data = common.convolutional(input_data, (1, 1, 1024, 512))
input_data = common.convolutional(input_data, (3, 3, 512, 1024))
input_data = common.convolutional(input_data, (1, 1, 1024, 512))
conv_lobj_branch = common.convolutional(input_data, (3, 3, 512, 1024))
conv_lbbox = common.convolutional(conv_lobj_branch,
(1, 1, 1024, 3 * (80 + 5)),
activate=False,
bn=False)
input_data = common.convolutional(input_data, (1, 1, 512, 256))
input_data = common.upsample(input_data)
input_data = tf.concat([input_data, route_2], axis=-1)
input_data = common.convolutional(input_data, (1, 1, 768, 256))
input_data = common.convolutional(input_data, (3, 3, 256, 512))
input_data = common.convolutional(input_data, (1, 1, 512, 256))
input_data = common.convolutional(input_data, (3, 3, 256, 512))
input_data = common.convolutional(input_data, (1, 1, 512, 256))
conv_mobj_branch = common.convolutional(input_data, (3, 3, 256, 512))
conv_mbbox = common.convolutional(conv_mobj_branch,
(1, 1, 512, 3 * (80 + 5)),
activate=False,
bn=False)
input_data = common.convolutional(input_data, (1, 1, 256, 128))
input_data = common.upsample(input_data)
input_data = tf.concat([input_data, route_1], axis=-1)
input_data = common.convolutional(input_data, (1, 1, 384, 128))
input_data = common.convolutional(input_data, (3, 3, 128, 256))
input_data = common.convolutional(input_data, (1, 1, 256, 128))
input_data = common.convolutional(input_data, (3, 3, 128, 256))
input_data = common.convolutional(input_data, (1, 1, 256, 128))
conv_sobj_branch = common.convolutional(input_data, (3, 3, 128, 256))
conv_sbbox = common.convolutional(conv_sobj_branch,
(1, 1, 256, 3 * (80 + 5)),
activate=False,
bn=False)
return conv_lbbox, conv_mbbox, conv_sbbox
def __build_nework(self, backbone):
route_1 = backbone.od_layer1
route_2 = backbone.od_layer2
input_data = backbone.od_layer4
input_data = common.convolutional(input_data, (1, 1, 512, 512), self.trainable, 'od_conv52')
input_data = common.convolutional(input_data, (3, 3, 512, 1024), self.trainable, 'od_conv53')
input_data = common.convolutional(input_data, (1, 1, 1024, 512), self.trainable, 'od_conv54')
input_data = common.convolutional(input_data, (3, 3, 512, 1024), self.trainable, 'od_conv55')
input_data = common.convolutional(input_data, (3, 3, 1024, 512), self.trainable, 'od_conv56')
conv_lobj_branch = common.convolutional(input_data, (3, 3, 512, 1024), self.trainable, name='od_conv_lobj_branch')
conv_lbbox = common.convolutional(conv_lobj_branch, (1, 1, 1024, 3*(self.num_class + 5)),
trainable=self.trainable, name='od_conv_lbbox', activate=False, bn=False)
input_data = common.convolutional(input_data, (1, 1, 512, 256), self.trainable, 'od_conv57')
input_data = common.upsample(input_data, name='od_upsample0', method=self.upsample_method)
with tf.variable_scope('route_1'):
input_data = tf.concat([input_data, route_2], axis=-1)
input_data = common.convolutional(input_data, (1, 1, 512, 256), self.trainable, 'od_conv58')
input_data = common.convolutional(input_data, (3, 3, 256, 512), self.trainable, 'od_conv59')
input_data = common.convolutional(input_data, (1, 1, 512, 256), self.trainable, 'od_conv60')
input_data = common.convolutional(input_data, (1, 1, 256, 512), self.trainable, 'od_conv61')
input_data = common.convolutional(input_data, (3, 3, 512, 256), self.trainable, 'od_conv62')
conv_mobj_branch = common.convolutional(input_data, (3, 3, 256, 512), self.trainable, name='od_conv_mobj_branch' )
conv_mbbox = common.convolutional(conv_mobj_branch, (1, 1, 512, 3*(self.num_class + 5)),
trainable=self.trainable, name='od_conv_mbbox', activate=False, bn=False)
input_data = common.convolutional(input_data, (1, 1, 256, 128), self.trainable, 'od_conv63')
input_data = common.upsample(input_data, name='od_upsample1', method=self.upsample_method)
with tf.variable_scope('route_2'):
input_data = tf.concat([input_data, route_1], axis=-1)
input_data = common.convolutional(input_data, (1, 1, 256, 128), self.trainable, 'od_conv64')
input_data = common.convolutional(input_data, (3, 3, 128, 256), self.trainable, 'od_conv65')
input_data = common.convolutional(input_data, (1, 1, 256, 128), self.trainable, 'od_conv66')
input_data = common.convolutional(input_data, (3, 3, 128, 256), self.trainable, 'od_conv67')
input_data = common.convolutional(input_data, (1, 1, 256, 128), self.trainable, 'od_conv68')
conv_sobj_branch = common.convolutional(input_data, (3, 3, 128, 256), self.trainable, name='od_conv_sobj_branch')
conv_sbbox = common.convolutional(conv_sobj_branch, (1, 1, 256, 3*(self.num_class + 5)),
trainable=self.trainable, name='od_conv_sbbox', activate=False, bn=False)
return conv_lbbox, conv_mbbox, conv_sbbox
def __build_nework(self, input_data):
route_1, route_2, input_data = backbone.darknet53(input_data, self.trainable)
input_data = common.convolutional(input_data, (1, 1, 1024, 512), self.trainable, 'conv52')
input_data = common.convolutional(input_data, (3, 3, 512, 1024), self.trainable, 'conv53')
input_data = common.convolutional(input_data, (1, 1, 1024, 512), self.trainable, 'conv54')
input_data = common.convolutional(input_data, (3, 3, 512, 1024), self.trainable, 'conv55')
input_data = common.convolutional(input_data, (1, 1, 1024, 512), self.trainable, 'conv56')
conv_lobj_branch = common.convolutional(input_data, (3, 3, 512, 1024), self.trainable, name='conv_lobj_branch')
conv_lbbox = common.convolutional(conv_lobj_branch, (1, 1, 1024, self.anchor_per_scale*(self.num_class + 5)),
trainable=self.trainable, name='conv_lbbox', activate=False, bn=False)
input_data = common.convolutional(input_data, (1, 1, 512, 256), self.trainable, 'conv57')
input_data = common.upsample(input_data, name='upsample0', method=self.upsample_method)
with tf.variable_scope('route_1'):
input_data = tf.concat([input_data, route_2], axis=-1)
input_data = common.convolutional(input_data, (1, 1, 768, 256), self.trainable, 'conv58')
input_data = common.convolutional(input_data, (3, 3, 256, 512), self.trainable, 'conv59')
input_data = common.convolutional(input_data, (1, 1, 512, 256), self.trainable, 'conv60')
input_data = common.convolutional(input_data, (3, 3, 256, 512), self.trainable, 'conv61')
input_data = common.convolutional(input_data, (1, 1, 512, 256), self.trainable, 'conv62')
conv_mobj_branch = common.convolutional(input_data, (3, 3, 256, 512), self.trainable, name='conv_mobj_branch' )
conv_mbbox = common.convolutional(conv_mobj_branch, (1, 1, 512, self.anchor_per_scale*(self.num_class + 5)),
trainable=self.trainable, name='conv_mbbox', activate=False, bn=False)
input_data = common.convolutional(input_data, (1, 1, 256, 128), self.trainable, 'conv63')
input_data = common.upsample(input_data, name='upsample1', method=self.upsample_method)
with tf.variable_scope('route_2'):
input_data = tf.concat([input_data, route_1], axis=-1)
input_data = common.convolutional(input_data, (1, 1, 384, 128), self.trainable, 'conv64')
input_data = common.convolutional(input_data, (3, 3, 128, 256), self.trainable, 'conv65')
input_data = common.convolutional(input_data, (1, 1, 256, 128), self.trainable, 'conv66')
input_data = common.convolutional(input_data, (3, 3, 128, 256), self.trainable, 'conv67')
input_data = common.convolutional(input_data, (1, 1, 256, 128), self.trainable, 'conv68')
conv_sobj_branch = common.convolutional(input_data, (3, 3, 128, 256), self.trainable, name='conv_sobj_branch')
conv_sbbox = common.convolutional(conv_sobj_branch, (1, 1, 256, self.anchor_per_scale*(self.num_class + 5)),
trainable=self.trainable, name='conv_sbbox', activate=False, bn=False)
return conv_lbbox, conv_mbbox, conv_sbbox
def upsample_2_wrapper(route_1, input_data):
input_data = common.upsample(input_data,
name="upsample1",
method=self.upsample_method,
precision=self.precision)
with tf.variable_scope("route_2"):
input_data = tf.concat([input_data, route_1], axis=-1)
return {"input_data": input_data}
def __build_nework(self, input_data):
route_8_layer, route_13_layer, input_data = backbone.darknet19(
input_data, self.trainable)
with tf.variable_scope('yolo1'):
conv_lobj_branch = common.convolutional(input_data,
(3, 3, 256, 512),
self.trainable,
name='conv_lobj_branch')
conv_lbbox = common.convolutional(
conv_lobj_branch, (1, 1, 512, 3 * (self.num_class + 5)),
trainable=self.trainable,
name='conv_lbbox',
activate=False,
bn=False)
with tf.variable_scope('route_1'):
input_data = route_13_layer
input_data = common.convolutional(input_data,
filters_shape=(3, 3, 256, 128),
trainable=self.trainable,
name='conv18')
input_data = common.upsample(input_data,
method=self.upsample_method,
name='upsample19')
with tf.variable_scope('route_2'):
input_data = tf.concat([input_data, route_8_layer],
axis=-1,
name='route20')
with tf.variable_scope('yolo2'):
conv_mobj_branch = common.convolutional(input_data,
filters_shape=(1, 1, 384,
256),
trainable=self.trainable,
name='conv21')
conv_mbbox = common.convolutional(
conv_mobj_branch,
filters_shape=(1, 1, 256, 3 * (self.num_class + 5)),
trainable=self.trainable,
name='conv22',
activate=False,
bn=False)
return conv_lbbox, conv_mbbox
def YOLOv3_tiny(input_layer, NUM_CLASS):
route_1, conv = backbone.darknet53_tiny(input_layer)
conv = common.convolutional(conv, (1, 1, 1024, 256))
conv_lobj_branch = common.convolutional(conv, (3, 3, 256, 512))
conv_lbbox = common.convolutional(conv_lobj_branch, (1, 1, 512, 3 * (NUM_CLASS + 5)), activate=False, bn=False)
conv = common.convolutional(conv, (1, 1, 256, 128))
conv = common.upsample(conv)
conv = tf.concat([conv, route_1], axis=-1)
conv_mobj_branch = common.convolutional(conv, (3, 3, 128, 256))
conv_mbbox = common.convolutional(conv_mobj_branch, (1, 1, 256, 3 * (NUM_CLASS + 5)), activate=False, bn=False)
return [conv_mbbox, conv_lbbox]
def __build_nework(self, input_data):
# backbone
input_data = common.convolutional(input_data, filters_shape=(3, 3, 3, 16), self.trainable, name='conv0')
input_data = tf.nn.max_pool(input_data, ksize=[1, 2, 2, 1], strides=[1, 2, 2, 1], padding='SAME')
input_data = common.convolutional(input_data, filters_shape=(3, 3, 16, 32), self.trainable, name='conv1')
input_data = tf.nn.max_pool(input_data, ksize=[1, 2, 2, 1], strides=[1, 2, 2, 1], padding='SAME')
input_data = common.convolutional(input_data, filters_shape=(3, 3, 32, 64), self.trainable, name='conv2')
input_data = tf.nn.max_pool(input_data, ksize=[1, 2, 2, 1], strides=[1, 2, 2, 1], padding='SAME')
input_data = common.convolutional(input_data, filters_shape=(3, 3, 64, 128), self.trainable, name='conv3')
input_data = tf.nn.max_pool(input_data, ksize=[1, 2, 2, 1], strides=[1, 2, 2, 1], padding='SAME')
input_data = common.convolutional(input_data, filters_shape=(3, 3, 128, 256), self.trainable, name='conv4')
route = input_data
input_data = tf.nn.max_pool(input_data, ksize=[1, 2, 2, 1], strides=[1, 2, 2, 1], padding='SAME')
input_data = common.convolutional(input_data, filters_shape=(3, 3, 256, 512), self.trainable, name='conv5')
input_data = tf.nn.max_pool(input_data, ksize=[1, 2, 2, 1], strides=[1, 2, 2, 1], padding='SAME')
input_data = common.convolutional(input_data, filters_shape=(3, 3, 512, 1024), self.trainable, name='conv6')
# end
input_data = common.convolutional(input_data, filters_shape=(1, 1, 1024, 256), self.trainable, name='conv7')
conv_lobj_branch = common.convolutional(input_data, (3, 3, 256, 512), self.trainable, name='conv_lobj_branch')
conv_lbbox = common.convolutional(conv_lobj_branch, (1, 1, 512, 3 * (self.num_class + 5)),
trainable=self.trainable, name='conv_lbbox', activate=False, bn=False)
input_data = common.convolutional(input_data, (1, 1, 256, 128), self.trainable, 'conv8')
input_data = common.upsample(input_data, name='upsample0', method=self.upsample_method)
with tf.variable_scope('route'):
input_data = tf.concat([input_data, route], axis=-1)
conv_mobj_branch = common.convolutional(input_data, (3, 3, 384, 256), self.trainable, name='conv_mobj_branch' )
conv_mbbox = common.convolutional(conv_mobj_branch, (1, 1, 256, 3 * (self.num_class + 5)),
trainable=self.trainable, name='conv_mbbox', activate=False, bn=False)
return conv_lbbox, conv_mbbox
def YOLOV3_TINY(input_layer):
#call darknet_tiny
route_1, route_2 = backbone.darknet_tiny(input_data=input_layer)
#call yolov3_tiny_block
route_3, mobj_branch = YOLOV3_TINY_block(route_2, 256)
#get feature map for 13x13x512 tensor
conv_mbbox = common.conv2d(mobj_branch,
filter_shape=(1, 1, 512, 3 * (NUM_CLASS + 5)),
activativation=False,
Bn=False)
route_3 = common.conv2d(input_layer=route_3, filter_shape=(1, 1, 256, 128))
route_3 = common.upsample(input_layer=route_3)
sobj_branch = tf.concat([route_3, route_1], axis=-1)
conv_sbbox = common.conv2d(sobj_branch,
filter_shape=(1, 1, 256, 3 * (NUM_CLASS + 5)),
activativation=False,
Bn=False)
return conv_sbbox, conv_mbbox
def test_upsample():
input_value = np.random.random((2, 224, 224, 3)).astype(np.float16)
# nearest_neighborhood function that run on cpu:
input_data = tf.constant(input_value)
input_shape = tf.shape(input_data)
upsample_cpu = tf.image.resize_nearest_neighbor(
input_data, (input_shape[1] * 2, input_shape[2] * 2))
with tf.Session() as sess:
cpu_output = sess.run(upsample_cpu)
tf.reset_default_graph()
# upsample that run on IPU:
config = ipu.config.IPUConfig()
config.auto_select_ipus = 1
config.configure_ipu_system()
with ipu.scopes.ipu_scope('/device:IPU:0'):
input_data = tf.constant(input_value)
upsample_ipu = upsample(input_data, "test_upsmaple", method="resize")
with tf.Session(config=tf.ConfigProto()) as sess:
ipu_output = sess.run(upsample_ipu)
assert np.max(cpu_output - ipu_output) <= 0.0001
def __build_nework(self, input_data):
"""
:param input_data: tf.placeholder(dtype=tf.float32, name='input_data')
:return: conv_lbbox [batch, 52, 52, 3*(self.num_class + 5)], conv_mbbox [batch, 26, 26, 3*(self.num_class + 5)], conv_sbbox [batch, 13, 13, 3*(self.num_class + 5)]
"""
route_1, route_2, input_data = backbone.darknet53(
input_data, self.trainable)
input_data = common.convolutional(input_data, (1, 1, 1024, 512),
self.trainable, 'conv52')
input_data = common.convolutional(input_data, (3, 3, 512, 1024),
self.trainable, 'conv53')
input_data = common.convolutional(input_data, (1, 1, 1024, 512),
self.trainable, 'conv54')
input_data = common.convolutional(input_data, (3, 3, 512, 1024),
self.trainable, 'conv55')
input_data = common.convolutional(input_data, (1, 1, 1024, 512),
self.trainable, 'conv56')
conv_lobj_branch = common.convolutional(input_data, (3, 3, 512, 1024),
self.trainable,
name='conv_lobj_branch')
conv_lbbox = common.convolutional(
conv_lobj_branch, (1, 1, 1024, 3 * (self.num_class + 5)),
trainable=self.trainable,
name='conv_lbbox',
activate=False,
bn=False)
input_data = common.convolutional(input_data, (1, 1, 512, 256),
self.trainable, 'conv57')
input_data = common.upsample(input_data,
name='upsample0',
method=self.upsample_method)
with tf.variable_scope('route_1'):
input_data = tf.concat([input_data, route_2], axis=-1)
input_data = common.convolutional(input_data, (1, 1, 768, 256),
self.trainable, 'conv58')
input_data = common.convolutional(input_data, (3, 3, 256, 512),
self.trainable, 'conv59')
input_data = common.convolutional(input_data, (1, 1, 512, 256),
self.trainable, 'conv60')
input_data = common.convolutional(input_data, (3, 3, 256, 512),
self.trainable, 'conv61')
input_data = common.convolutional(input_data, (1, 1, 512, 256),
self.trainable, 'conv62')
conv_mobj_branch = common.convolutional(input_data, (3, 3, 256, 512),
self.trainable,
name='conv_mobj_branch')
conv_mbbox = common.convolutional(
conv_mobj_branch, (1, 1, 512, 3 * (self.num_class + 5)),
trainable=self.trainable,
name='conv_mbbox',
activate=False,
bn=False)
input_data = common.convolutional(input_data, (1, 1, 256, 128),
self.trainable, 'conv63')
input_data = common.upsample(input_data,
name='upsample1',
method=self.upsample_method)
with tf.variable_scope('route_2'):
input_data = tf.concat([input_data, route_1], axis=-1)
input_data = common.convolutional(input_data, (1, 1, 384, 128),
self.trainable, 'conv64')
input_data = common.convolutional(input_data, (3, 3, 128, 256),
self.trainable, 'conv65')
input_data = common.convolutional(input_data, (1, 1, 256, 128),
self.trainable, 'conv66')
input_data = common.convolutional(input_data, (3, 3, 128, 256),
self.trainable, 'conv67')
input_data = common.convolutional(input_data, (1, 1, 256, 128),
self.trainable, 'conv68')
conv_sobj_branch = common.convolutional(input_data, (3, 3, 128, 256),
self.trainable,
name='conv_sobj_branch')
conv_sbbox = common.convolutional(
conv_sobj_branch, (1, 1, 256, 3 * (self.num_class + 5)),
trainable=self.trainable,
name='conv_sbbox',
activate=False,
bn=False)
return conv_lbbox, conv_mbbox, conv_sbbox # 13,26,52
def __build_nework(self, input_data):
# route1 2 3 input data 分别为用于后面concat的值 即 128残差后用于最大特征图concat的 route1 以此类推
route_1, route_2, input_data = backbone.darknet53(
input_data, self.trainable)
# 经过全部残差的 output 先经过五次DBL操作
input_data = common.convolutional(input_data, (1, 1, 1024, 512),
self.trainable, 'conv52')
input_data = common.convolutional(input_data, (3, 3, 512, 1024),
self.trainable, 'conv53')
input_data = common.convolutional(input_data, (1, 1, 1024, 512),
self.trainable, 'conv54')
input_data = common.convolutional(input_data, (3, 3, 512, 1024),
self.trainable, 'conv55')
input_data = common.convolutional(input_data, (1, 1, 1024, 512),
self.trainable, 'conv56')
conv_lobj_branch = common.convolutional(input_data, (3, 3, 512, 1024),
self.trainable,
name='conv_lobj_branch')
conv_lbbox = common.convolutional(
conv_lobj_branch, (1, 1, 1024, 3 * (self.num_class + 5)),
trainable=self.trainable,
name='conv_lbbox',
activate=False,
bn=False)
input_data = common.convolutional(input_data, (1, 1, 512, 256),
self.trainable, 'conv57')
input_data = common.upsample(input_data,
name='upsample0',
method=self.upsample_method)
with tf.variable_scope('route_1'):
input_data = tf.concat([input_data, route_2], axis=-1)
input_data = common.convolutional(input_data, (1, 1, 768, 256),
self.trainable, 'conv58')
input_data = common.convolutional(input_data, (3, 3, 256, 512),
self.trainable, 'conv59')
input_data = common.convolutional(input_data, (1, 1, 512, 256),
self.trainable, 'conv60')
input_data = common.convolutional(input_data, (3, 3, 256, 512),
self.trainable, 'conv61')
input_data = common.convolutional(input_data, (1, 1, 512, 256),
self.trainable, 'conv62')
conv_mobj_branch = common.convolutional(input_data, (3, 3, 256, 512),
self.trainable,
name='conv_mobj_branch')
conv_mbbox = common.convolutional(
conv_mobj_branch, (1, 1, 512, 3 * (self.num_class + 5)),
trainable=self.trainable,
name='conv_mbbox',
activate=False,
bn=False)
input_data = common.convolutional(input_data, (1, 1, 256, 128),
self.trainable, 'conv63')
input_data = common.upsample(input_data,
name='upsample1',
method=self.upsample_method)
with tf.variable_scope('route_2'):
input_data = tf.concat([input_data, route_1], axis=-1)
input_data = common.convolutional(input_data, (1, 1, 384, 128),
self.trainable, 'conv64')
input_data = common.convolutional(input_data, (3, 3, 128, 256),
self.trainable, 'conv65')
input_data = common.convolutional(input_data, (1, 1, 256, 128),
self.trainable, 'conv66')
input_data = common.convolutional(input_data, (3, 3, 128, 256),
self.trainable, 'conv67')
input_data = common.convolutional(input_data, (1, 1, 256, 128),
self.trainable, 'conv68')
conv_sobj_branch = common.convolutional(input_data, (3, 3, 128, 256),
self.trainable,
name='conv_sobj_branch')
conv_sbbox = common.convolutional(
conv_sobj_branch, (1, 1, 256, 3 * (self.num_class + 5)),
trainable=self.trainable,
name='conv_sbbox',
activate=False,
bn=False)
# 返回三种大小的检测框 13*13*255 26*26 52*52
return conv_lbbox, conv_mbbox, conv_sbbox
def __build_nework(self, input_data):
# mobilenet v1 backbone
# input_data = tf.layers.conv2d(input_data,
# filters=32,
# kernel_size=(3, 3),
# strides=(2,2),
# padding = 'same',
# activation = tf.nn.relu,
# name = 'conv1'
# )
# input_data = tf.layers.batch_normalization(input_data, beta_initializer=tf.zeros_initializer(),
# gamma_initializer=tf.ones_initializer(),
# moving_mean_initializer=tf.zeros_initializer(),
# moving_variance_initializer=tf.ones_initializer(), training=self.trainable,
# name='bn')
'''
input_data = common.convolutional(input_data, (3, 3, 3, 32), self.trainable, 'conv0')
input_data = self.separable_conv_block(input=input_data, dw_filter=(3,3,32,1),output_channel=64,
strides=(1,1,1,1), name="spearable_1")
input_data = self.separable_conv_block(input=input_data, dw_filter=(3, 3, 64, 1), output_channel=128,
strides=(1, 2, 2, 1), name="spearable_2")
input_data = self.separable_conv_block(input=input_data, dw_filter=(3,3,128,1),output_channel=128,
strides=(1,1,1,1), name="spearable_3")
input_data = self.separable_conv_block(input=input_data, dw_filter=(3, 3, 128, 1), output_channel=256,
strides=(1, 2, 2, 1), name="spearable_4")
input_data = self.separable_conv_block(input=input_data, dw_filter=(3, 3, 256, 1), output_channel=256,
strides=(1, 1, 1, 1), name="spearable_5")
with tf.variable_scope("spearable_6"):
dw_weight = tf.get_variable(name='dw_filter', dtype=tf.float32, trainable=True,
shape=(3, 3, 256, 1), initializer=tf.random_normal_initializer(stddev=0.01))
input_data = tf.nn.depthwise_conv2d(input=input_data, filter=dw_weight, strides=(1, 2, 2, 1), padding="SAME", name='Conv/dw')
input_data = tf.layers.batch_normalization(input_data, beta_initializer=tf.zeros_initializer(),
gamma_initializer=tf.ones_initializer(),
moving_mean_initializer=tf.zeros_initializer(),
moving_variance_initializer=tf.ones_initializer(), training=self.trainable,
name='dw/bn')
route = tf.nn.leaky_relu(input_data,0.1)
weight = tf.get_variable(name='weight', dtype=tf.float32, trainable=True,
shape=(1, 1, 256, 512), initializer=tf.random_normal_initializer(stddev=0.01))
input_data = tf.nn.conv2d(input=route, filter=weight, strides=[1, 1, 1, 1], padding="SAME",name="conv/s1")
input_data = tf.layers.batch_normalization(input_data, beta_initializer=tf.zeros_initializer(),
gamma_initializer=tf.ones_initializer(),
moving_mean_initializer=tf.zeros_initializer(),
moving_variance_initializer=tf.ones_initializer(),
training=self.trainable,
name='pt/bn')
input_data = tf.nn.leaky_relu(input_data,0.1)
for i in range(5):
input_data = self.separable_conv_block(input=input_data, dw_filter=(3, 3, 512, 1), output_channel=512,
strides=(1, 1, 1, 1), name="spearable_%d" % (i+ 7))
input_data = self.separable_conv_block(input=input_data, dw_filter=(3, 3, 512, 1), output_channel=1024,
strides=(1, 2, 2, 1), name="spearable_12")
input_data = self.separable_conv_block(input=input_data, dw_filter=(3, 3, 1024, 1), output_channel=1024,
strides=(1, 1, 1, 1), name="spearable_13")
'''
# mobilenet backbone end
# yolo tiny backbone
input_data = common.convolutional(input_data, (3, 3, 3, 16),
self.trainable, 'conv0')
input_data = tf.nn.max_pool(input_data,
ksize=[1, 2, 2, 1],
strides=[1, 2, 2, 1],
padding='SAME')
input_data = common.convolutional(input_data, (3, 3, 16, 32),
self.trainable, 'conv1')
input_data = tf.nn.max_pool(input_data,
ksize=[1, 2, 2, 1],
strides=[1, 2, 2, 1],
padding='SAME')
input_data = common.convolutional(input_data, (3, 3, 32, 64),
self.trainable, 'conv2')
input_data = tf.nn.max_pool(input_data,
ksize=[1, 2, 2, 1],
strides=[1, 2, 2, 1],
padding='SAME')
input_data = common.convolutional(input_data, (3, 3, 64, 128),
self.trainable, 'conv3')
input_data = tf.nn.max_pool(input_data,
ksize=[1, 2, 2, 1],
strides=[1, 2, 2, 1],
padding='SAME')
input_data = common.convolutional(input_data, (3, 3, 128, 256),
self.trainable, 'conv4')
route = input_data
input_data = tf.nn.max_pool(input_data,
ksize=[1, 2, 2, 1],
strides=[1, 2, 2, 1],
padding='SAME')
input_data = common.convolutional(input_data, (3, 3, 256, 512),
self.trainable, 'conv5')
input_data = tf.nn.max_pool(input_data,
ksize=[1, 2, 2, 1],
strides=[1, 1, 1, 1],
padding='SAME')
input_data = common.convolutional(input_data, (3, 3, 512, 1024),
self.trainable, 'conv6')
# end
# yolo detect layer
input_data = common.convolutional(input_data, (1, 1, 1024, 256),
self.trainable, 'conv7')
conv_lobj_branch = common.convolutional(input_data, (3, 3, 256, 512),
self.trainable,
'conv_lobj_branch')
conv_lbbox = common.convolutional(
conv_lobj_branch, (1, 1, 512, 3 * (self.num_class + 5)),
self.trainable,
'conv_lbbox',
activate=False,
bn=False)
input_data = common.convolutional(input_data, (1, 1, 256, 128),
self.trainable, 'conv8')
input_data = common.upsample(input_data,
name='upsample0',
method=self.upsample_method)
with tf.variable_scope('route_1'):
input_data = tf.concat([input_data, route], axis=-1)
conv_mobj_branch = common.convolutional(input_data, (3, 3, 384, 256),
self.trainable,
'conv_mobj_branch')
conv_mbbox = common.convolutional(
conv_mobj_branch, (1, 1, 256, 3 * (self.num_class + 5)),
self.trainable,
'conv_mbbox',
activate=False,
bn=False)
# yolo detect layer end
# yolo detect layer DSC
'''
input_data = self.separable_conv_block(input=input_data, dw_filter=(1, 1, 1024, 1),output_channel=256,
strides=(1,1,1,1), name="conv7")
conv_lobj_branch = self.separable_conv_block(input=input_data, dw_filter=(3, 3, 256, 1),output_channel=512,
strides=(1,1,1,1), name="conv_lobj_branch")
conv_lbbox = self.separable_conv_block(input=conv_lobj_branch, dw_filter=(1, 1, 512, 1),output_channel=3 * (self.num_class + 5),
strides=(1,1,1,1), name="conv_lbbox")
input_data = self.separable_conv_block(input=input_data, dw_filter=(1, 1, 256, 1),output_channel=128,
strides=(1,1,1,1), name="conv8")
input_data = common.upsample(input_data, name='upsample0', method=self.upsample_method)
with tf.variable_scope('route_1'):
input_data = tf.concat([input_data, route], axis=-1)
conv_mobj_branch = self.separable_conv_block(input=input_data, dw_filter=(3, 3, 384, 1),output_channel=256,
strides=(1,1,1,1), name="conv_mobj_branch")
conv_mbbox = self.separable_conv_block(input=conv_mobj_branch, dw_filter=(1, 1, 256, 1),output_channel=3 * (self.num_class + 5),
strides=(1,1,1,1), name="conv_mbbox")
'''
return conv_lbbox, conv_mbbox
def __build_nework(self, input_data):
route_1, route_2, input_data = backbone.darknet53(
input_data, self.trainable)
conv_lobj_branch = common.EP(input_data,
189,
462,
name='EP5',
stride=1)
conv_lbbox = common.convolutional(
conv_lobj_branch, (1, 1, 462, 3 * (self.num_class + 5)),
trainable=self.trainable,
name='conv_lbbox',
activate=False,
bn=False)
input_data = common.convolutional(input_data, (1, 1, 189, 105),
self.trainable, 'conv5')
input_data = common.upsample(input_data,
name='upsample0',
method=self.upsample_method)
input_data = tf.concat([input_data, route_2], axis=-1)
input_data = common.PEP(input_data,
430,
325,
113,
name='PEP17',
stride=1)
input_data = common.PEP(input_data,
325,
207,
99,
name='PEP18',
stride=1)
input_data = common.convolutional(input_data, (1, 1, 207, 98),
self.trainable, 'conv6')
conv_mobj_branch = common.EP(input_data, 98, 183, name='EP6', stride=1)
conv_mbbox = common.convolutional(
conv_mobj_branch, (1, 1, 183, 3 * (self.num_class + 5)),
trainable=self.trainable,
name='conv_mbbox',
activate=False,
bn=False)
input_data = common.convolutional(input_data, (1, 1, 98, 47),
self.trainable, 'conv7')
input_data = common.upsample(input_data,
name='upsample1',
method=self.upsample_method)
input_data = tf.concat([input_data, route_1], axis=-1)
input_data = common.PEP(input_data,
197,
122,
58,
name='PEP19',
stride=1)
input_data = common.PEP(input_data,
122,
87,
52,
name='PEP20',
stride=1)
conv_sobj_branch = common.PEP(input_data,
87,
93,
47,
name='PEP21',
stride=1)
conv_sbbox = common.convolutional(conv_sobj_branch,
(1, 1, 93, 3 * (self.num_class + 5)),
trainable=self.trainable,
name='conv_sbbox',
activate=False,
bn=False)
return conv_lbbox, conv_mbbox, conv_sbbox
def __build_tiny_3l(self, input_data):
input_data = input_data / 255.
input_data = common.convolutional(input_data,
filters_shape=(3, 3, 3, 16),
trainable=self.trainable,
name='conv0')
input_data = tf.nn.max_pool(input_data,
ksize=(1, 2, 2, 1),
strides=(1, 2, 2, 1),
padding='SAME',
name='max_pool0')
for i in range(5):
input_data = common.convolutional(
input_data,
filters_shape=(3, 3, 16 * pow(2, i), 16 * pow(2, i + 1)),
trainable=self.trainable,
name='conv{}'.format(i + 1))
if i == 2:
route_0 = input_data
if i == 3:
route_1 = input_data
if i == 4:
input_data = tf.nn.max_pool(input_data,
ksize=(1, 2, 2, 1),
strides=(1, 1, 1, 1),
padding='SAME',
name='max_pool{}'.format(i + 1))
else:
input_data = tf.nn.max_pool(input_data,
ksize=(1, 2, 2, 1),
strides=(1, 2, 2, 1),
padding='SAME',
name='max_pool{}'.format(i + 1))
input_data = common.convolutional(input_data,
filters_shape=(3, 3, 512, 1024),
trainable=self.trainable,
name='conv6')
input_data = common.convolutional(input_data,
filters_shape=(1, 1, 1024, 256),
trainable=self.trainable,
name='conv7')
conv_lobj_branch = common.convolutional(input_data, (3, 3, 256, 512),
self.trainable,
name='conv_lobj_branch')
conv_lbbox = common.convolutional(
conv_lobj_branch, (1, 1, 512, 3 * (self.num_class + 5)),
trainable=self.trainable,
name='conv_lbbox',
activate=False,
bn=False)
input_data = common.convolutional(input_data,
filters_shape=(1, 1, 256, 128),
trainable=self.trainable,
name='conv8')
input_data = common.upsample(input_data,
name='upsample0',
method=self.upsample_method)
with tf.variable_scope('route_1'):
input_data = tf.concat([input_data, route_1], axis=-1)
conv_mobj_branch = common.convolutional(input_data, (3, 3, 384, 256),
self.trainable,
name='conv_mobj_branch')
conv_mbbox = common.convolutional(
conv_mobj_branch, (1, 1, 256, 3 * (self.num_class + 5)),
trainable=self.trainable,
name='conv_mbbox',
activate=False,
bn=False)
input_data = common.convolutional(conv_mobj_branch,
filters_shape=(1, 1, 256, 128),
trainable=self.trainable,
name='conv9')
input_data = common.upsample(input_data,
name='upsample1',
method=self.upsample_method)
with tf.variable_scope('route_0'):
input_data = tf.concat([input_data, route_0], axis=-1)
conv_sobj_branch = common.convolutional(input_data, (3, 3, 256, 128),
self.trainable,
name='conv_sobj_branch')
conv_sbbox = common.convolutional(
conv_sobj_branch, (1, 1, 128, 3 * (self.num_class + 5)),
trainable=self.trainable,
name='conv_sbbox',
activate=False,
bn=False)
return conv_lbbox, conv_mbbox, conv_sbbox
def YOLOv5(input_layer, NUM_CLASS, scale_v5):
V5_STR = utils.get_V5STR(scale_v5)
print("YOLOv5 structure parameters:",V5_STR)
N_start_next, route_1, route_2, route_3, G1, G2 = backbone.Focus_cspdarknet53(input_layer,V5_STR)
###Detection branch###
G3, G4 = route_1, route_2
route_3 = common.convolutional(route_3, (1, 1, V5_STR[9][0], V5_STR[10][0]),activate_type="SiLU",name=[0,N_start_next,N_start_next])
R1 = route_3
route_3 = common.upsample(route_3,activate_type="SiLU",name=[0,N_start_next+1,N_start_next+1])
route_3 = tf.concat([route_2, route_3], axis=-1)
route_3 = common.C3(route_3, V5_STR[10][0]*2, V5_STR[11][0], V5_STR[11][1], cut1=False,C3_name=[0,N_start_next+2])
N_start_next = N_start_next+5+2*V5_STR[11][1]
route_3 = common.convolutional(route_3, (1, 1, V5_STR[11][0], V5_STR[12][0]),activate_type="SiLU", name=[0,N_start_next,N_start_next])
R2 = route_3
route_3 = common.upsample(route_3,activate_type="SiLU", name=[0,N_start_next+1,N_start_next+1])
route_3 = tf.concat([route_1, route_3], axis=-1)
route_3 = common.C3(route_3, V5_STR[12][0]*2, V5_STR[13][0], V5_STR[13][1], cut1=False,C3_name=[0,N_start_next+2])
N_start_next = N_start_next+5+2*V5_STR[13][1]
conv_sbbox = common.convolutional(route_3, (1, 1, V5_STR[13][0], 3 * (NUM_CLASS + 5)), activate=False, bn=False, name=[0,N_start_next,N_start_next])
route_3 = common.convolutional(route_3, (3, 3, V5_STR[13][0], V5_STR[14][0]), downsample=True, activate_type="SiLU", name=[0,N_start_next+1,N_start_next+1])
route_3 = tf.concat([R2, route_3], axis=-1)
route_3 = common.C3(route_3, V5_STR[14][0]*2, V5_STR[15][0], V5_STR[15][1], cut1=False,C3_name=[0,N_start_next+2])
N_start_next = N_start_next+5+2*V5_STR[15][1]
conv_mbbox = common.convolutional(route_3, (1, 1, V5_STR[15][0], 3 * (NUM_CLASS + 5)), activate=False, bn=False, name=[0,N_start_next,N_start_next])
route_3 = common.convolutional(route_3, (3, 3, V5_STR[15][0], V5_STR[16][0]), downsample=True, activate_type="SiLU", name=[0,N_start_next+1,N_start_next+1])
route_3 = tf.concat([R1, route_3], axis=-1)
route_3 = common.C3(route_3, V5_STR[16][0]*2, V5_STR[17][0],V5_STR[17][1], cut1=False, C3_name=[0,N_start_next+2])
N_start_next = N_start_next+5+2*V5_STR[17][1]
conv_lbbox = common.convolutional(route_3, (1, 1, V5_STR[17][0], 3 * (NUM_CLASS + 5)), activate=False, bn=False,name=[0,N_start_next,N_start_next])
###Deblurring branch###
out2 = common.convolutional(G2, (1, 1, 128, 128), activate=False, bn=False, name=[1,N_start_next+1,N_start_next+1])
out2 = common.CSP_CBAM(out2, activate_type1="SiLU", CB_name=[1,N_start_next+2])
N_start_next = N_start_next+12
out3 = common.convolutional(G3, (1, 1, 256, 128), activate=False, bn=False, name=[1,N_start_next,N_start_next])
out3 = common.CSP_CBAM(out3, activate_type1="SiLU", CB_name=[1,N_start_next+1])
N_start_next = N_start_next+11
out2 = tf.concat([common.upsample(out3,activate_type="SiLU",name=[1,N_start_next,N_start_next]), out2], axis=-1)
out2 = common.convolutional(common.upsample(out2,activate_type="SiLU",name=[1,N_start_next+1,N_start_next+1]), (3, 3, 256, 64),activate_type="SiLU",name=[1,N_start_next+2,N_start_next+2])
N_start_next = N_start_next+3
out4 = common.convolutional(G4, (1, 1, 512, 128), activate=False, bn=False,name=[1,N_start_next,N_start_next])
out4 = common.CSP_CBAM(out4, activate_type1="SiLU", CB_name=[1,N_start_next+1])
N_start_next = N_start_next+11
out3 = tf.concat([common.upsample(out4,activate_type="SiLU",name=[1,N_start_next,N_start_next]), out3], axis=-1)
out3 = common.convolutional(common.upsample(out3,activate_type="SiLU",name=[1,N_start_next+1,N_start_next+1]), (3, 3, 256, 128),activate_type="SiLU",name=[1,N_start_next+2,N_start_next+2])
out3 = common.convolutional(common.upsample(out3,activate_type="SiLU",name=[1,N_start_next+3,N_start_next+3]), (3, 3, 128, 64),activate_type="SiLU",name=[1,N_start_next+4,N_start_next+4])
N_start_next = N_start_next+5
out4 = common.convolutional(common.upsample(out4,activate_type="SiLU",name=[1,N_start_next,N_start_next]), (3, 3, 128, 128),activate_type="SiLU",name=[1,N_start_next+1,N_start_next+1])
out4 = common.convolutional(common.upsample(out4,activate_type="SiLU",name=[1,N_start_next+2,N_start_next+2]), (3, 3, 128, 64),activate_type="SiLU",name=[1,N_start_next+3,N_start_next+3])
out4 = common.convolutional(common.upsample(out4,activate_type="SiLU",name=[1,N_start_next+4,N_start_next+4]), (3, 3, 64, 64),activate_type="SiLU",name=[1,N_start_next+5,N_start_next+5])
N_start_next = N_start_next+6
concat = tf.concat([out2, out3, out4], axis=-1)
concat = common.convolutional(concat, (3, 3, 64*3, 128),activate_type="SiLU",name=[1,N_start_next,N_start_next])
out1 = common.convolutional(G1, (1, 1, 64, 128),activate=False, bn=False,name=[1,N_start_next+1,N_start_next+1])
out1 = common.CSP_CBAM(out1, activate_type1="SiLU", CB_name=[1,N_start_next+2])
N_start_next = N_start_next+12
out1 = tf.concat([out1, concat], axis=-1)
out1 = common.convolutional(common.upsample(out1,activate_type="SiLU",name=[1,N_start_next,N_start_next]), (3, 3, 256, 32),activate_type="SiLU",name=[1,N_start_next+1,N_start_next+1])
out1 = common.convolutional(out1, (1, 1, 32, 3), activate=False, bn=False,name=[1,N_start_next+2,N_start_next+2])
out1 = tf.math.sigmoid(out1) + input_layer
result = tf.clip_by_value(out1, clip_value_min=0, clip_value_max=1)
return [conv_sbbox, conv_mbbox, conv_lbbox, result]
def YOLOv4(input_layer, NUM_CLASS):
print('[INFO][core.yolov4.YOLOv4] Initializing YOLOv4 architecture')
route_1, route_2, conv = backbone.cspdarknet53(input_layer)
print('[INFO][core.yolov4.YOLOv4] cspdarknet53 backbone initialized')
route = conv
conv = common.convolutional(conv, (1, 1, 512, 256))
conv = common.upsample(conv)
route_2 = common.convolutional(route_2, (1, 1, 512, 256))
conv = tf.concat([route_2, conv], axis=-1)
conv = common.convolutional(conv, (1, 1, 512, 256))
conv = common.convolutional(conv, (3, 3, 256, 512))
conv = common.convolutional(conv, (1, 1, 512, 256))
conv = common.convolutional(conv, (3, 3, 256, 512))
conv = common.convolutional(conv, (1, 1, 512, 256))
route_2 = conv
conv = common.convolutional(conv, (1, 1, 256, 128))
conv = common.upsample(conv)
route_1 = common.convolutional(route_1, (1, 1, 256, 128))
conv = tf.concat([route_1, conv], axis=-1)
conv = common.convolutional(conv, (1, 1, 256, 128))
conv = common.convolutional(conv, (3, 3, 128, 256))
conv = common.convolutional(conv, (1, 1, 256, 128))
conv = common.convolutional(conv, (3, 3, 128, 256))
conv = common.convolutional(conv, (1, 1, 256, 128))
route_1 = conv
conv = common.convolutional(conv, (3, 3, 128, 256))
conv_sbbox = common.convolutional(conv, (1, 1, 256, 3 * (NUM_CLASS + 5)),
activate=False,
bn=False)
print(f'[DEBUG][core.yolov4.YOLOv4] conv_ssbox : {conv_sbbox}')
conv = common.convolutional(route_1, (3, 3, 128, 256), downsample=True)
conv = tf.concat([conv, route_2], axis=-1)
conv = common.convolutional(conv, (1, 1, 512, 256))
conv = common.convolutional(conv, (3, 3, 256, 512))
conv = common.convolutional(conv, (1, 1, 512, 256))
conv = common.convolutional(conv, (3, 3, 256, 512))
conv = common.convolutional(conv, (1, 1, 512, 256))
route_2 = conv
conv = common.convolutional(conv, (3, 3, 256, 512))
conv_mbbox = common.convolutional(conv, (1, 1, 512, 3 * (NUM_CLASS + 5)),
activate=False,
bn=False)
print(f'[DEBUG][core.yolov4.YOLOv4] conv_mbbox : {conv_mbbox}')
conv = common.convolutional(route_2, (3, 3, 256, 512), downsample=True)
conv = tf.concat([conv, route], axis=-1)
conv = common.convolutional(conv, (1, 1, 1024, 512))
conv = common.convolutional(conv, (3, 3, 512, 1024))
conv = common.convolutional(conv, (1, 1, 1024, 512))
conv = common.convolutional(conv, (3, 3, 512, 1024))
conv = common.convolutional(conv, (1, 1, 1024, 512))
conv = common.convolutional(conv, (3, 3, 512, 1024))
conv_lbbox = common.convolutional(conv, (1, 1, 1024, 3 * (NUM_CLASS + 5)),
activate=False,
bn=False)
print(f'[DEBUG][core.yolov4.YOLOv4] conv_lbbox : {conv_lbbox}')
print(f'[INFO][core.yolov4.YOLOv4] YOLOv4 architecture initialized')
return [conv_sbbox, conv_mbbox, conv_lbbox]
def __build_nework(self, input_data):
net_name = "YOLOV3"
with tf.variable_scope(net_name):
route_1, route_2, input_data = backbone.MobilenetV2(
input_data, self.trainable)
input_data = common.convolutional(input_data, (1, 1, 1280, 512),
self.trainable, 'conv52')
input_data = separable_conv(name="conv53",
input_data=input_data,
input_c=512,
output_c=1024,
training=self.trainable)
input_data = common.convolutional(input_data, (1, 1, 1024, 512),
self.trainable, 'conv54')
input_data = separable_conv(name="conv55",
input_data=input_data,
input_c=512,
output_c=1024,
training=self.trainable)
input_data = common.convolutional(input_data, (1, 1, 1024, 512),
self.trainable, 'conv56')
conv_lobj_branch = separable_conv(name="conv_lobj_branch",
input_data=input_data,
input_c=512,
output_c=1024,
training=self.trainable)
conv_lbbox = common.convolutional(
conv_lobj_branch, (1, 1, 1024, 3 * (self.num_class + 5)),
trainable=self.trainable,
name='conv_lbbox',
activate=False,
bn=False)
input_data = common.convolutional(input_data, (1, 1, 512, 256),
self.trainable, 'conv57')
input_data = common.upsample(input_data,
name='upsample0',
method=self.upsample_method)
with tf.variable_scope('route_1'):
input_data = tf.concat([input_data, route_2], axis=-1)
input_data = common.convolutional(input_data, (1, 1, 352, 256),
self.trainable, 'conv58')
input_data = separable_conv(name="conv59",
input_data=input_data,
input_c=256,
output_c=512,
training=self.trainable)
input_data = common.convolutional(input_data, (1, 1, 512, 256),
self.trainable, 'conv60')
input_data = separable_conv(name="conv61",
input_data=input_data,
input_c=256,
output_c=512,
training=self.trainable)
input_data = common.convolutional(input_data, (1, 1, 512, 256),
self.trainable, 'conv62')
conv_mobj_branch = separable_conv(name="conv_mobj_branch",
input_data=input_data,
input_c=256,
output_c=512,
training=self.trainable)
conv_mbbox = common.convolutional(
conv_mobj_branch, (1, 1, 512, 3 * (self.num_class + 5)),
trainable=self.trainable,
name='conv_mbbox',
activate=False,
bn=False)
input_data = common.convolutional(input_data, (1, 1, 256, 128),
self.trainable, 'conv63')
input_data = common.upsample(input_data,
name='upsample1',
method=self.upsample_method)
with tf.variable_scope('route_2'):
input_data = tf.concat([input_data, route_1], axis=-1)
input_data = common.convolutional(input_data, (1, 1, 160, 128),
self.trainable, 'conv64')
input_data = separable_conv(name="conv_65",
input_data=input_data,
input_c=128,
output_c=256,
training=self.trainable)
input_data = common.convolutional(input_data, (1, 1, 256, 128),
self.trainable, 'conv66')
input_data = separable_conv(name="conv_67",
input_data=input_data,
input_c=128,
output_c=256,
training=self.trainable)
input_data = common.convolutional(input_data, (1, 1, 256, 128),
self.trainable, 'conv68')
conv_sobj_branch = separable_conv(name="conv_sobj_branch",
input_data=input_data,
input_c=128,
output_c=256,
training=self.trainable)
conv_sbbox = common.convolutional(
conv_sobj_branch, (1, 1, 256, 3 * (self.num_class + 5)),
trainable=self.trainable,
name='conv_sbbox',
activate=False,
bn=False)
for var in tf.global_variables(net_name):
tf.add_to_collection(net_name, var)
return conv_lbbox, conv_mbbox, conv_sbbox
def __build_nework(self, input_data):
input_data = common.convolutional(input_data, (3, 3, 1, 16),
self.trainable,
name='conv0',
downsample=True)
input_data = common.convolutional(input_data, (3, 3, 16, 32),
self.trainable,
name='conv1',
downsample=True)
input_data = common.convolutional(input_data, (3, 3, 32, 32),
self.trainable,
name='conv2',
downsample=True)
input_data = common.convolutional(input_data, (3, 3, 32, 32),
self.trainable,
name='conv3',
downsample=True)
input_data = common.convolutional(input_data, (3, 3, 32, 64),
self.trainable,
name='conv4',
downsample=False) #16
route_1 = input_data
input_data = common.convolutional(input_data, (3, 3, 64, 64),
self.trainable,
name='conv5',
downsample=True) #32
input_data = common.convolutional(input_data, (3, 3, 64, 128),
self.trainable,
name='conv6',
downsample=False)
input_data = common.convolutional(input_data, (1, 1, 128, 64),
self.trainable,
name='conv7',
downsample=False)
conv_lobj_branch = common.convolutional(input_data, (3, 3, 64, 128),
self.trainable,
name='conv_lobj_branch')
conv_lbbox = common.convolutional(
conv_lobj_branch, (1, 1, 128, 3 * (self.num_class + 5)),
trainable=self.trainable,
name='conv_lbbox',
activate=False,
bn=False)
input_data = common.convolutional(input_data, (1, 1, 64, 32),
self.trainable,
name='conv8',
downsample=False)
input_data = common.upsample(input_data,
name='upsample0',
method=self.upsample_method) #16
route_2 = input_data
input_data = tf.concat([route_1, route_2], axis=-1)
conv_mobj_branch = common.convolutional(input_data, (3, 3, 96, 32),
self.trainable,
name='conv_mobj_branch')
conv_mbbox = common.convolutional(conv_mobj_branch,
(1, 1, 32, 3 * (self.num_class + 5)),
trainable=self.trainable,
name='conv_mbbox',
activate=False,
bn=False)
return conv_lbbox, conv_mbbox
def __build_nework(self, input_data):
route_1, route_2, input_data = backbone.darknet53(
input_data, self.trainable)
# input_data = common.convolutional(input_data, (1, 1, 1024, 512), self.trainable, 'conv52')
# input_data = common.convolutional(input_data, (3, 3, 512, 1024), self.trainable, 'conv53')
# input_data = common.convolutional(input_data, (1, 1, 1024, 512), self.trainable, 'conv54')
# input_data = common.convolutional(input_data, (3, 3, 512, 1024), self.trainable, 'conv55')
# input_data = common.convolutional(input_data, (1, 1, 1024, 512), self.trainable, 'conv56')
#
# conv_lobj_branch = common.convolutional(input_data, (3, 3, 512, 1024), self.trainable, name='conv_lobj_branch')
# conv_lbbox = common.convolutional(conv_lobj_branch, (1, 1, 1024, 3 * (self.num_class + 5)),
# trainable=self.trainable, name='conv_lbbox', activate=False, bn=False)
#
# input_data = common.convolutional(input_data, (1, 1, 512, 256), self.trainable, 'conv57')
input_data = common.upsample(input_data,
name='upsample0',
method=self.upsample_method)
with tf.variable_scope('route_1'):
input_data = tf.concat([input_data, route_2],
axis=-1) # 1024+512=1536
# input_data = common.convolutional(input_data, (1, 1, 768, 256), self.trainable, 'conv58')
# input_data = common.convolutional(input_data, (3, 3, 256, 512), self.trainable, 'conv59')
# input_data = common.convolutional(input_data, (1, 1, 512, 256), self.trainable, 'conv60')
# input_data = common.convolutional(input_data, (3, 3, 256, 512), self.trainable, 'conv61')
# input_data = common.convolutional(input_data, (1, 1, 512, 256), self.trainable, 'conv62')
#
# conv_mobj_branch = common.convolutional(input_data, (3, 3, 256, 512), self.trainable, name='conv_mobj_branch')
# conv_mbbox = common.convolutional(conv_mobj_branch, (1, 1, 512, 3 * (self.num_class + 5)),
# trainable=self.trainable, name='conv_mbbox', activate=False, bn=False)
#
# input_data = common.convolutional(input_data, (1, 1, 256, 128), self.trainable, 'conv63')
input_data = common.upsample(input_data,
name='upsample1',
method=self.upsample_method)
with tf.variable_scope('route_2'):
input_data = tf.concat([input_data, route_1],
axis=-1) # 1536+256=1792
# input_data = common.convolutional(input_data, (1, 1, 384, 128), self.trainable, 'conv64')
# input_data = common.convolutional(input_data, (3, 3, 128, 256), self.trainable, 'conv65')
# input_data = common.convolutional(input_data, (1, 1, 256, 128), self.trainable, 'conv66')
# input_data = common.convolutional(input_data, (3, 3, 128, 256), self.trainable, 'conv67')
# input_data = common.convolutional(input_data, (1, 1, 256, 128), self.trainable, 'conv68')
# conv_sobj_branch = common.convolutional(input_data, (3, 3, 128, 256), self.trainable, name='conv_sobj_branch')
# conv_s = common.convolutional(conv_sobj_branch, (1, 1, 256, 3 * (self.num_class + 5)),
# trainable=self.trainable, name='conv_sbbox', activate=False, bn=False)
input_data = common.convolutional(input_data,
filters_shape=(3, 3, 1792, 256),
trainable=True,
name='conv_s_conv',
downsample=True) # 26*26*256
input_data = common.convolutional(input_data,
filters_shape=(1, 1, 256,
self.num_class),
trainable=True,
name='conv_s_conv1',
downsample=True) # 13*13*19
conv_s = tf.layers.flatten(input_data)
conv_s = tf.reshape(conv_s, (-1, 13 * 13 * self.num_class))
conv_s = tf.layers.dense(
conv_s,
units=1024,
activation=tf.nn.leaky_relu,
kernel_regularizer=tf.contrib.layers.l2_regularizer(0.1))
conv_s = tf.layers.dropout(conv_s, rate=0.5)
conv_s = tf.layers.dense(
conv_s,
units=512,
activation=tf.nn.leaky_relu,
kernel_regularizer=tf.contrib.layers.l2_regularizer(0.1))
conv_s = tf.layers.dropout(conv_s, rate=0.5)
conv_s = tf.layers.dense(conv_s,
units=self.num_class,
name='conv_l',
activation=tf.nn.softmax)
return conv_s
def __build_nework(self, input_data):
print('__build_nework self.net_type=', self.net_type)
assert self.net_type in [
'darknet53', 'mobilenetv2', 'mobilenetv3', 'mobilenetv3_small'
]
if self.net_type == 'darknet53':
route_1, route_2, input_data = backbone.darknet53(
input_data, self.trainable)
input_data = common.convolutional(input_data, (1, 1, 1024, 512),
self.trainable, 'conv52')
input_data = common.convolutional(input_data, (3, 3, 512, 1024),
self.trainable, 'conv53')
input_data = common.convolutional(input_data, (1, 1, 1024, 512),
self.trainable, 'conv54')
input_data = common.convolutional(input_data, (3, 3, 512, 1024),
self.trainable, 'conv55')
input_data = common.convolutional(input_data, (1, 1, 1024, 512),
self.trainable, 'conv56')
conv_lobj_branch = common.convolutional(input_data,
(3, 3, 512, 1024),
self.trainable,
name='conv_lobj_branch')
conv_lbbox = common.convolutional(
conv_lobj_branch, (1, 1, 1024, 3 * (self.num_class + 5)),
trainable=self.trainable,
name='conv_lbbox',
activate=False,
bn=False)
input_data = common.convolutional(input_data, (1, 1, 512, 256),
self.trainable, 'conv57')
input_data = common.upsample(input_data,
name='upsample0',
method=self.upsample_method)
with tf.variable_scope('route_1'):
input_data = tf.concat([input_data, route_2], axis=-1)
input_data = common.convolutional(input_data, (1, 1, 768, 256),
self.trainable, 'conv58')
input_data = common.convolutional(input_data, (3, 3, 256, 512),
self.trainable, 'conv59')
input_data = common.convolutional(input_data, (1, 1, 512, 256),
self.trainable, 'conv60')
input_data = common.convolutional(input_data, (3, 3, 256, 512),
self.trainable, 'conv61')
input_data = common.convolutional(input_data, (1, 1, 512, 256),
self.trainable, 'conv62')
conv_mobj_branch = common.convolutional(input_data,
(3, 3, 256, 512),
self.trainable,
name='conv_mobj_branch')
conv_mbbox = common.convolutional(
conv_mobj_branch, (1, 1, 512, 3 * (self.num_class + 5)),
trainable=self.trainable,
name='conv_mbbox',
activate=False,
bn=False)
input_data = common.convolutional(input_data, (1, 1, 256, 128),
self.trainable, 'conv63')
input_data = common.upsample(input_data,
name='upsample1',
method=self.upsample_method)
with tf.variable_scope('route_2'):
input_data = tf.concat([input_data, route_1], axis=-1)
input_data = common.convolutional(input_data, (1, 1, 384, 128),
self.trainable, 'conv64')
input_data = common.convolutional(input_data, (3, 3, 128, 256),
self.trainable, 'conv65')
input_data = common.convolutional(input_data, (1, 1, 256, 128),
self.trainable, 'conv66')
input_data = common.convolutional(input_data, (3, 3, 128, 256),
self.trainable, 'conv67')
input_data = common.convolutional(input_data, (1, 1, 256, 128),
self.trainable, 'conv68')
conv_sobj_branch = common.convolutional(input_data,
(3, 3, 128, 256),
self.trainable,
name='conv_sobj_branch')
conv_sbbox = common.convolutional(
conv_sobj_branch, (1, 1, 256, 3 * (self.num_class + 5)),
trainable=self.trainable,
name='conv_sbbox',
activate=False,
bn=False)
return conv_lbbox, conv_mbbox, conv_sbbox
elif self.net_type == 'mobilenetv2':
route_1, route_2, input_data = backbone.mobilenetv2(
input_data, self.trainable)
input_data = common.convolutional(input_data, (1, 1, 320, 160),
self.trainable, 'conv18')
input_data = common.convolutional(input_data, (3, 3, 160, 320),
self.trainable, 'conv19')
input_data = common.convolutional(input_data, (1, 1, 320, 160),
self.trainable, 'conv20')
conv_lobj_branch = common.convolutional(input_data,
(3, 3, 160, 320),
self.trainable,
name='conv_lobj_branch')
conv_lbbox = common.convolutional(
conv_lobj_branch, (1, 1, 320, 3 * (self.num_class + 5)),
trainable=self.trainable,
name='conv_lbbox',
activate=False,
bn=False)
input_data = common.convolutional(input_data, (1, 1, 160, 80),
self.trainable, 'conv21')
input_data = common.upsample(input_data,
name='upsample0',
method=self.upsample_method)
with tf.variable_scope('route_1'):
input_data = tf.concat([input_data, route_2], axis=-1)
input_data = common.convolutional(input_data, (1, 1, 176, 80),
self.trainable, 'conv22')
input_data = common.convolutional(input_data, (3, 3, 80, 160),
self.trainable, 'conv23')
input_data = common.convolutional(input_data, (1, 1, 160, 80),
self.trainable, 'conv24')
conv_mobj_branch = common.convolutional(input_data,
(3, 3, 80, 160),
self.trainable,
name='conv_mobj_branch')
conv_mbbox = common.convolutional(
conv_mobj_branch, (1, 1, 160, 3 * (self.num_class + 5)),
trainable=self.trainable,
name='conv_mbbox',
activate=False,
bn=False)
input_data = common.convolutional(input_data, (1, 1, 80, 40),
self.trainable, 'conv25')
input_data = common.upsample(input_data,
name='upsample1',
method=self.upsample_method)
with tf.variable_scope('route_2'):
input_data = tf.concat([input_data, route_1], axis=-1)
input_data = common.convolutional(input_data, (1, 1, 72, 40),
self.trainable, 'conv26')
input_data = common.convolutional(input_data, (3, 3, 40, 80),
self.trainable, 'conv27')
input_data = common.convolutional(input_data, (1, 1, 80, 40),
self.trainable, 'conv28')
conv_sobj_branch = common.convolutional(input_data, (3, 3, 40, 80),
self.trainable,
name='conv_sobj_branch')
conv_sbbox = common.convolutional(
conv_sobj_branch, (1, 1, 80, 3 * (self.num_class + 5)),
trainable=self.trainable,
name='conv_sbbox',
activate=False,
bn=False)
return conv_lbbox, conv_mbbox, conv_sbbox
elif self.net_type == 'mobilenetv3':
route_1, route_2, input_data = backbone.mobilenetv3(
input_data, self.trainable)
input_data = common.convolutional(input_data, (1, 1, 160, 80),
self.trainable, 'conv8')
input_data = common.convolutional(input_data, (3, 3, 80, 160),
self.trainable, 'conv9')
input_data = common.convolutional(input_data, (1, 1, 160, 80),
self.trainable, 'conv10')
conv_lobj_branch = common.convolutional(input_data,
(3, 3, 80, 160),
self.trainable,
name='conv_lobj_branch')
conv_lbbox = common.convolutional(
conv_lobj_branch, (1, 1, 160, 3 * (self.num_class + 5)),
trainable=self.trainable,
name='conv_lbbox',
activate=False,
bn=False)
input_data = common.convolutional(input_data, (1, 1, 80, 40),
self.trainable, 'conv11')
input_data = common.upsample(input_data,
name='upsample0',
method=self.upsample_method)
with tf.variable_scope('route_1'):
input_data = tf.concat([input_data, route_2], axis=-1)
input_data = common.convolutional(input_data, (1, 1, 152, 40),
self.trainable, 'conv12')
input_data = common.convolutional(input_data, (3, 3, 40, 80),
self.trainable, 'conv13')
input_data = common.convolutional(input_data, (1, 1, 80, 40),
self.trainable, 'conv14')
conv_mobj_branch = common.convolutional(input_data, (3, 3, 40, 80),
self.trainable,
name='conv_mobj_branch')
conv_mbbox = common.convolutional(
conv_mobj_branch, (1, 1, 80, 3 * (self.num_class + 5)),
trainable=self.trainable,
name='conv_mbbox',
activate=False,
bn=False)
input_data = common.convolutional(input_data, (1, 1, 40, 20),
self.trainable, 'conv15')
input_data = common.upsample(input_data,
name='upsample1',
method=self.upsample_method)
with tf.variable_scope('route_2'):
input_data = tf.concat([input_data, route_1], axis=-1)
input_data = common.convolutional(input_data, (1, 1, 60, 20),
self.trainable, 'conv16')
input_data = common.convolutional(input_data, (3, 3, 20, 40),
self.trainable, 'conv17')
input_data = common.convolutional(input_data, (1, 1, 40, 20),
self.trainable, 'conv18')
conv_sobj_branch = common.convolutional(input_data, (3, 3, 20, 40),
self.trainable,
name='conv_sobj_branch')
conv_sbbox = common.convolutional(
conv_sobj_branch, (1, 1, 40, 3 * (self.num_class + 5)),
trainable=self.trainable,
name='conv_sbbox',
activate=False,
bn=False)
return conv_lbbox, conv_mbbox, conv_sbbox
elif self.net_type == 'mobilenetv3_small':
route_1, route_2, input_data = backbone.mobilenetv3_small(
input_data, self.trainable)
input_data = common.convolutional(input_data, (1, 1, 96, 80),
self.trainable, 'conv7')
input_data = common.convolutional(input_data, (3, 3, 80, 160),
self.trainable, 'conv8')
input_data = common.convolutional(input_data, (1, 1, 160, 80),
self.trainable, 'conv9')
conv_lobj_branch = common.convolutional(input_data,
(3, 3, 80, 160),
self.trainable,
name='conv_lobj_branch')
conv_lbbox = common.convolutional(
conv_lobj_branch, (1, 1, 160, 3 * (self.num_class + 5)),
trainable=self.trainable,
name='conv_lbbox',
activate=False,
bn=False)
input_data = common.convolutional(input_data, (1, 1, 80, 40),
self.trainable, 'conv10')
input_data = common.upsample(input_data,
name='upsample0',
method=self.upsample_method)
with tf.variable_scope('route_1'):
input_data = tf.concat([input_data, route_2], axis=-1)
input_data = common.convolutional(input_data, (1, 1, 88, 40),
self.trainable, 'conv11')
input_data = common.convolutional(input_data, (3, 3, 40, 80),
self.trainable, 'conv12')
input_data = common.convolutional(input_data, (1, 1, 80, 40),
self.trainable, 'conv13')
conv_mobj_branch = common.convolutional(input_data, (3, 3, 40, 80),
self.trainable,
name='conv_mobj_branch')
conv_mbbox = common.convolutional(
conv_mobj_branch, (1, 1, 80, 3 * (self.num_class + 5)),
trainable=self.trainable,
name='conv_mbbox',
activate=False,
bn=False)
input_data = common.convolutional(input_data, (1, 1, 40, 20),
self.trainable, 'conv14')
input_data = common.upsample(input_data,
name='upsample1',
method=self.upsample_method)
with tf.variable_scope('route_2'):
input_data = tf.concat([input_data, route_1], axis=-1)
input_data = common.convolutional(input_data, (1, 1, 44, 20),
self.trainable, 'conv15')
input_data = common.convolutional(input_data, (3, 3, 20, 40),
self.trainable, 'conv16')
input_data = common.convolutional(input_data, (1, 1, 40, 20),
self.trainable, 'conv17')
conv_sobj_branch = common.convolutional(input_data, (3, 3, 20, 40),
self.trainable,
name='conv_sobj_branch')
conv_sbbox = common.convolutional(
conv_sobj_branch, (1, 1, 40, 3 * (self.num_class + 5)),
trainable=self.trainable,
name='conv_sbbox',
activate=False,
bn=False)
return conv_lbbox, conv_mbbox, conv_sbbox
else:
print('self.net_type=%s error' % self.net_type)
def object_detection_network(self, input_data):
route_1, route_2, input_data = backbone.darknet53(
input_data, self.trainable)
conv_output = common.convolutional(input_data, (1, 1, 1024, 512),
self.trainable, 'conv53')
conv_output = common.convolutional(conv_output, (3, 3, 512, 1024),
self.trainable, 'conv54')
conv_output = common.convolutional(conv_output, (1, 1, 1024, 512),
self.trainable, 'conv55')
conv_output = common.convolutional(conv_output, (3, 3, 512, 1024),
self.trainable, 'conv56')
conv_output = common.convolutional(conv_output, (1, 1, 1024, 512),
self.trainable, 'conv57')
conv_lobj_branch = common.convolutional(conv_output, (3, 3, 512, 1024),
self.trainable,
'conv_lobj_branch')
conv_lbbox = common.convolutional(
conv_lobj_branch, (1, 1, 1024, 3 * (self.num_class + 5)),
trainable=self.trainable,
activate=False,
bn=False,
name='conv_lbbox')
conv_output = common.convolutional(conv_output, (1, 1, 512, 256),
self.trainable, 'conv58')
conv_upsampled_1 = common.upsample(name='upsample1',
input_data=conv_output,
method=self.upsample_method)
with tf.variable_scope('route_1'):
conv_output = tf.concat([conv_upsampled_1, route_2], axis=-1)
conv_output = common.convolutional(conv_output, (1, 1, 768, 256),
self.trainable, 'conv59')
conv_output = common.convolutional(conv_output, (3, 3, 256, 512),
self.trainable, 'conv60')
conv_output = common.convolutional(conv_output, (1, 1, 512, 256),
self.trainable, 'conv61')
conv_output = common.convolutional(conv_output, (3, 3, 256, 512),
self.trainable, 'conv62')
conv_output = common.convolutional(conv_output, (1, 1, 512, 256),
self.trainable, 'conv63')
conv_mobj_branch = common.convolutional(conv_output,
(3, 3, 256, 512),
self.trainable,
'conv_mobj_branch')
conv_mbbox = common.convolutional(
conv_mobj_branch, (1, 1, 512, 3 * (self.num_class + 5)),
self.trainable,
activate=False,
bn=False,
name='conv_mbbox')
conv_output = common.convolutional(conv_output, (1, 1, 256, 128),
self.trainable, 'conv64')
conv_upsampled_2 = common.upsample(name='upsample2',
input_data=conv_output,
method=self.upsample_method)
with tf.variable_scope('route_2'):
conv_output = tf.concat([conv_upsampled_2, route_1], axis=-1)
conv_output = common.convolutional(conv_output, (1, 1, 384, 128),
self.trainable, 'conv65')
conv_output = common.convolutional(conv_output, (3, 3, 128, 256),
self.trainable, 'conv66')
conv_output = common.convolutional(conv_output, (1, 1, 256, 128),
self.trainable, 'conv67')
conv_output = common.convolutional(conv_output, (3, 3, 128, 256),
self.trainable, 'conv68')
conv_output = common.convolutional(conv_output, (1, 1, 256, 128),
self.trainable, 'conv69')
conv_sobj_branch = common.convolutional(conv_output,
(3, 3, 128, 256),
self.trainable, 'conv70')
conv_sbbox = common.convolutional(
conv_sobj_branch, (1, 1, 256, 3 * (self.num_class + 5)),
self.trainable,
activate=False,
bn=False,
name='conv_sbbox')
return conv_lbbox, conv_mbbox, conv_sbbox
def YOLOv3(input_layer):
if BACK_BONE == 'resnet18':
route_1, route_2, conv = backbone.resnet(input_layer, 18)
if BACK_BONE == 'resnet34':
route_1, route_2, conv = backbone.resnet(input_layer, 34)
if BACK_BONE == 'resnet50':
route_1, route_2, conv = backbone.resnet(input_layer, 50)
if BACK_BONE == 'resnet101':
route_1, route_2, conv = backbone.resnet(input_layer, 101)
if BACK_BONE == 'resnet152':
route_1, route_2, conv = backbone.resnet(input_layer, 152)
if BACK_BONE == 'darknet53':
route_1, route_2, conv = backbone.darknet53(input_layer)
conv = common.convolutional(conv, (1, 1, 1024, 512))
conv = common.convolutional(conv, (3, 3, 512, 1024))
conv = common.convolutional(conv, (1, 1, 1024, 512))
conv = common.convolutional(conv, (3, 3, 512, 1024))
conv = common.convolutional(conv, (1, 1, 1024, 512))
conv_lobj_branch = common.convolutional(conv, (3, 3, 512, 1024))
conv_lbbox = common.convolutional(conv_lobj_branch,
(1, 1, 1024, 3 * (NUM_CLASS + 5)),
activate=False,
bn=False)
conv = common.convolutional(conv, (1, 1, 512, 256))
conv = common.upsample(conv)
conv = tf.concat([conv, route_2], axis=-1)
conv = common.convolutional(conv, (1, 1, 768, 256))
conv = common.convolutional(conv, (3, 3, 256, 512))
conv = common.convolutional(conv, (1, 1, 512, 256))
conv = common.convolutional(conv, (3, 3, 256, 512))
conv = common.convolutional(conv, (1, 1, 512, 256))
conv_mobj_branch = common.convolutional(conv, (3, 3, 256, 512))
conv_mbbox = common.convolutional(conv_mobj_branch,
(1, 1, 512, 3 * (NUM_CLASS + 5)),
activate=False,
bn=False)
conv = common.convolutional(conv, (1, 1, 256, 128))
conv = common.upsample(conv)
conv = tf.concat([conv, route_1], axis=-1)
conv = common.convolutional(conv, (1, 1, 384, 128))
conv = common.convolutional(conv, (3, 3, 128, 256))
conv = common.convolutional(conv, (1, 1, 256, 128))
conv = common.convolutional(conv, (3, 3, 128, 256))
conv = common.convolutional(conv, (1, 1, 256, 128))
conv_sobj_branch = common.convolutional(conv, (3, 3, 128, 256))
conv_sbbox = common.convolutional(conv_sobj_branch,
(1, 1, 256, 3 * (NUM_CLASS + 5)),
activate=False,
bn=False)
return [conv_sbbox, conv_mbbox, conv_lbbox]
def yolo_v3_network(self, input_layer):
# 输入层进入 Darknet-53 网络后，得到了三个分支
middle_data_1, middle_data_2, conv = darknet53.darknet53(input_layer)
# 五次DBL(卷积，BN[批量规范化],激活函数)
conv = common.convolutional_Layer(conv, (1, 1, 1024, 512),
name='conv52_five_1')
conv = common.convolutional_Layer(conv, (3, 3, 512, 1024),
name='conv53_five_1')
conv = common.convolutional_Layer(conv, (1, 1, 1024, 512),
name='conv54_five_1')
conv = common.convolutional_Layer(conv, (3, 3, 512, 1024),
name='conv55_five_1')
conv = common.convolutional_Layer(conv, (1, 1, 1024, 512),
name='conv56_five_1')
# 第一尺度 ----------->conv_BIG_BOX 用于预测大尺寸物体，shape = [None, 13, 13, 255]
conv_first_obj_branch = common.convolutional_Layer(
conv, (3, 3, 512, 1024), name='conv_big_obj_branch')
conv_big_BOX = common.convolutional_Layer(
conv_first_obj_branch, (1, 1, 1024, 3 * (self.class_num + 5)),
activate=False,
bn=False,
name='conv_big_BOX')
# --------------------------------
conv = common.convolutional_Layer(conv, (1, 1, 512, 256),
name='conv57')
conv = common.upsample(conv, 'middle_box_upsample')
conv = tf.concat([conv, middle_data_2],
axis=-1) # axis,拼接方向，1y轴，2x轴，3...-1-->最后一维度
# 此刻的输入维度（512+256)
conv = common.convolutional_Layer(conv, (1, 1, 768, 256),
name='conv58_five_2')
conv = common.convolutional_Layer(conv, (3, 3, 256, 512),
name='conv59_five_2')
conv = common.convolutional_Layer(conv, (1, 1, 512, 256),
name='conv60_five_2')
conv = common.convolutional_Layer(conv, (3, 3, 256, 512),
name='conv61_five_2')
conv = common.convolutional_Layer(conv, (1, 1, 512, 256),
name='conv62_five_2')
# 第二尺度 ----------->conv_middle_BOX 用于预测中等尺寸物体，shape = [None, 26, 26, 255]
conv_second_obj_branch = common.convolutional_Layer(
conv, (3, 3, 256, 512), name='conv_mid_obj_branch')
conv_middle_BOX = common.convolutional_Layer(
conv_second_obj_branch, (1, 1, 512, 3 * (self.class_num + 5)),
activate=False,
bn=False,
name='conv_mid_BOX')
# --------------------------------
conv = common.convolutional_Layer(conv, (1, 1, 256, 128),
name='conv63')
conv = common.upsample(conv, 'smart_box_upsample')
conv = tf.concat([conv, middle_data_1], axis=-1)
# 此刻的输入维度（256+128）
conv = common.convolutional_Layer(conv, (1, 1, 384, 128),
name='conv64_five_3')
conv = common.convolutional_Layer(conv, (3, 3, 128, 256),
name='conv65_five_3')
conv = common.convolutional_Layer(conv, (1, 1, 256, 128),
name='conv66_five_3')
conv = common.convolutional_Layer(conv, (3, 3, 128, 256),
name='conv67_five_3')
conv = common.convolutional_Layer(conv, (1, 1, 256, 128),
name='conv68_five_3')
# 第三尺度 ----------->conv_smart_BOX 用于预测小尺寸物体，shape = [None, 52, 52, 255]
conv_third_obj_branch = common.convolutional_Layer(
conv, (3, 3, 128, 256), name='conv_smt_obj_branch')
conv_smart_BOX = common.convolutional_Layer(
conv_third_obj_branch, (1, 1, 256, 3 * (self.class_num + 5)),
activate=False,
bn=False,
name='conv_smt_BOX')
return conv_smart_BOX, conv_middle_BOX, conv_big_BOX