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 locaNet(input_layer):
input_data = common.convolutional(input_layer, (3, 3, cfg.INPUT_CHANNEL, 8)) # 320x224x8
input_data = common.convolutional(input_data, (3, 3, 8, 16), downsample=True) # 160x112x16
input_data = common.convolutional(input_data, (3, 3, 16, 32), downsample=True) # 80x56x32
input_data = common.convolutional(input_data, (3, 3, 32, 64), downsample=True) # 40x28x64
route_1 = input_data
input_data = common.convolutional(input_data, (3, 3, 64, 128), downsample=True) # 20x14x128
route_2 = input_data
input_data = common.convolutional(input_data, (3, 3, 128, 256), downsample=True) # 10x7x256
conv = common.convolutional(input_data, (1, 1, 256, 128)) # 10x7x128
conv = common.upsample(conv) # 20x14x128
conv = tf.keras.layers.Concatenate(axis=-1)([conv, route_2]) # 20x14x256
conv = common.convolutional(conv, (1, 1, 256, 128)) # 20x14x128
conv = common.upsample(conv) # 40x28x128
conv = tf.keras.layers.Concatenate(axis=-1)([conv, route_1]) # 40x28x192
conv_point = common.convolutional(input_data, (1, 1, 192, 2+NUM_CLASS), activate=False, bn=False) # 40x28x(2 + NUM_CLASS)
return conv_point
def YOLOv3(input_layer, NUM_CLASS):
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 yolo3(input_layer):
branch_1,branch_2,conv = backbone.backbone(input_layer)
conv = common.conv(conv,(1,512))
conv = common.conv(conv,(3,1024))
conv = common.conv(conv,(1,512))
conv = common.conv(conv,(3,1024))
conv = common.conv(conv,(1,512))
conv_l_branch = common.conv(conv,(3,1024))
l_output = common.conv(conv_l_branch,(1,3 * (NUM_CLASS + 5)),activation = False,bn = False)
conv = common.conv(conv,(1,256))
conv = common.upsample(conv,"resize")
conv = tf.concat([conv,branch_2],axis = -1)
conv = common.conv(conv,(1,256))
conv = common.conv(conv,(3,512))
conv = common.conv(conv,(1,256))
conv = common.conv(conv,(3,512))
conv = common.conv(conv,(1,256))
conv_m_branch = common.conv(conv,(3,512))
m_output = common.conv(conv_m_branch,(1,3 * (NUM_CLASS + 5)),activation = False,bn = False)
conv = common.conv(conv,(1,128))
conv = common.upsample(conv,"resize")
conv = tf.concat([conv,branch_1],axis = -1)
conv = common.conv(conv,(1,128))
conv = common.conv(conv,(3,256))
conv = common.conv(conv,(1,128))
conv = common.conv(conv,(3,256))
conv = common.conv(conv,(1,128))
conv_s_branch = common.conv(conv,(3,256))
s_output = common.conv(conv_s_branch,(1,3 * (NUM_CLASS + 5)),activation = False,bn = False)
return [s_output,m_output,l_output]