1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
|
from ..net import Net, Mode
from ..utils import LeakyReLU, average_endpoint_error, pad, antipad
from ..correlation import correlation
from ..downsample import downsample
import math
import tensorflow as tf
slim = tf.contrib.slim
class FlowNetC(Net):
def __init__(self, mode=Mode.TRAIN, debug=False):
super(FlowNetC, self).__init__(mode=mode, debug=debug)
def model(self, inputs, training_schedule, trainable=True):
_, height, width, _ = inputs['input_a'].shape.as_list()
with tf.variable_scope('FlowNetC'):
with slim.arg_scope([slim.conv2d, slim.conv2d_transpose],
# Only backprop this network if trainable
trainable=trainable,
# He (aka MSRA) weight initialization
weights_initializer=slim.variance_scaling_initializer(),
activation_fn=LeakyReLU,
# We will do our own padding to match the original Caffe code
padding='VALID'):
weights_regularizer = slim.l2_regularizer(training_schedule['weight_decay'])
with slim.arg_scope([slim.conv2d], weights_regularizer=weights_regularizer):
with slim.arg_scope([slim.conv2d], stride=2):
conv_a_1 = slim.conv2d(pad(inputs['input_a'], 3), 64, 7, scope='conv1')
conv_a_2 = slim.conv2d(pad(conv_a_1, 2), 128, 5, scope='conv2')
conv_a_3 = slim.conv2d(pad(conv_a_2, 2), 256, 5, scope='conv3')
conv_b_1 = slim.conv2d(pad(inputs['input_b'], 3),
64, 7, scope='conv1', reuse=True)
conv_b_2 = slim.conv2d(pad(conv_b_1, 2), 128, 5, scope='conv2', reuse=True)
conv_b_3 = slim.conv2d(pad(conv_b_2, 2), 256, 5, scope='conv3', reuse=True)
# Compute cross correlation with leaky relu activation
cc = correlation(conv_a_3, conv_b_3, 1, 20, 1, 2, 20)
cc_relu = LeakyReLU(cc)
# Combine cross correlation results with convolution of feature map A
netA_conv = slim.conv2d(conv_a_3, 32, 1, scope='conv_redir')
# Concatenate along the channels axis
net = tf.concat([netA_conv, cc_relu], axis=3)
conv3_1 = slim.conv2d(pad(net), 256, 3, scope='conv3_1')
with slim.arg_scope([slim.conv2d], num_outputs=512, kernel_size=3):
conv4 = slim.conv2d(pad(conv3_1), stride=2, scope='conv4')
conv4_1 = slim.conv2d(pad(conv4), scope='conv4_1')
conv5 = slim.conv2d(pad(conv4_1), stride=2, scope='conv5')
conv5_1 = slim.conv2d(pad(conv5), scope='conv5_1')
conv6 = slim.conv2d(pad(conv5_1), 1024, 3, stride=2, scope='conv6')
conv6_1 = slim.conv2d(pad(conv6), 1024, 3, scope='conv6_1')
""" START: Refinement Network """
with slim.arg_scope([slim.conv2d_transpose], biases_initializer=None):
predict_flow6 = slim.conv2d(pad(conv6_1), 2, 3,
scope='predict_flow6',
activation_fn=None)
deconv5 = antipad(slim.conv2d_transpose(conv6_1, 512, 4,
stride=2,
scope='deconv5'))
upsample_flow6to5 = antipad(slim.conv2d_transpose(predict_flow6, 2, 4,
stride=2,
scope='upsample_flow6to5',
activation_fn=None))
concat5 = tf.concat([conv5_1, deconv5, upsample_flow6to5], axis=3)
predict_flow5 = slim.conv2d(pad(concat5), 2, 3,
scope='predict_flow5',
activation_fn=None)
deconv4 = antipad(slim.conv2d_transpose(concat5, 256, 4,
stride=2,
scope='deconv4'))
upsample_flow5to4 = antipad(slim.conv2d_transpose(predict_flow5, 2, 4,
stride=2,
scope='upsample_flow5to4',
activation_fn=None))
concat4 = tf.concat([conv4_1, deconv4, upsample_flow5to4], axis=3)
predict_flow4 = slim.conv2d(pad(concat4), 2, 3,
scope='predict_flow4',
activation_fn=None)
deconv3 = antipad(slim.conv2d_transpose(concat4, 128, 4,
stride=2,
scope='deconv3'))
upsample_flow4to3 = antipad(slim.conv2d_transpose(predict_flow4, 2, 4,
stride=2,
scope='upsample_flow4to3',
activation_fn=None))
concat3 = tf.concat([conv3_1, deconv3, upsample_flow4to3], axis=3)
predict_flow3 = slim.conv2d(pad(concat3), 2, 3,
scope='predict_flow3',
activation_fn=None)
deconv2 = antipad(slim.conv2d_transpose(concat3, 64, 4,
stride=2,
scope='deconv2'))
upsample_flow3to2 = antipad(slim.conv2d_transpose(predict_flow3, 2, 4,
stride=2,
scope='upsample_flow3to2',
activation_fn=None))
concat2 = tf.concat([conv_a_2, deconv2, upsample_flow3to2], axis=3)
predict_flow2 = slim.conv2d(pad(concat2), 2, 3,
scope='predict_flow2',
activation_fn=None)
""" END: Refinement Network """
flow = predict_flow2 * 20.0
# TODO: Look at Accum (train) or Resample (deploy) to see if we need to do something different
flow = tf.image.resize_bilinear(flow,
tf.stack([height, width]),
align_corners=True)
return {
'predict_flow6': predict_flow6,
'predict_flow5': predict_flow5,
'predict_flow4': predict_flow4,
'predict_flow3': predict_flow3,
'predict_flow2': predict_flow2,
'flow': flow,
}
def loss(self, flow, predictions):
flow = flow * 0.05
losses = []
INPUT_HEIGHT, INPUT_WIDTH = float(flow.shape[1].value), float(flow.shape[2].value)
# L2 loss between predict_flow6, blob23 (weighted w/ 0.32)
predict_flow6 = predictions['predict_flow6']
size = [predict_flow6.shape[1], predict_flow6.shape[2]]
downsampled_flow6 = downsample(flow, size)
losses.append(average_endpoint_error(downsampled_flow6, predict_flow6))
# L2 loss between predict_flow5, blob28 (weighted w/ 0.08)
predict_flow5 = predictions['predict_flow5']
size = [predict_flow5.shape[1], predict_flow5.shape[2]]
downsampled_flow5 = downsample(flow, size)
losses.append(average_endpoint_error(downsampled_flow5, predict_flow5))
# L2 loss between predict_flow4, blob33 (weighted w/ 0.02)
predict_flow4 = predictions['predict_flow4']
size = [predict_flow4.shape[1], predict_flow4.shape[2]]
downsampled_flow4 = downsample(flow, size)
losses.append(average_endpoint_error(downsampled_flow4, predict_flow4))
# L2 loss between predict_flow3, blob38 (weighted w/ 0.01)
predict_flow3 = predictions['predict_flow3']
size = [predict_flow3.shape[1], predict_flow3.shape[2]]
downsampled_flow3 = downsample(flow, size)
losses.append(average_endpoint_error(downsampled_flow3, predict_flow3))
# L2 loss between predict_flow2, blob43 (weighted w/ 0.005)
predict_flow2 = predictions['predict_flow2']
size = [predict_flow2.shape[1], predict_flow2.shape[2]]
downsampled_flow2 = downsample(flow, size)
losses.append(average_endpoint_error(downsampled_flow2, predict_flow2))
loss = tf.losses.compute_weighted_loss(losses, [0.32, 0.08, 0.02, 0.01, 0.005])
# Return the 'total' loss: loss fns + regularization terms defined in the model
return tf.losses.get_total_loss()
|
