/
losses.py
217 lines (171 loc) · 8.37 KB
/
losses.py
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
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
"""
Copyright 2017-2018 Fizyr (https://fizyr.com)
Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.
"""
# import keras
from tensorflow import keras
import tensorflow as tf
def focal(alpha=0.25, gamma=2.0):
"""
Create a functor for computing the focal loss.
Args
alpha: Scale the focal weight with alpha.
gamma: Take the power of the focal weight with gamma.
Returns
A functor that computes the focal loss using the alpha and gamma.
"""
def _focal(y_true, y_pred):
"""
Compute the focal loss given the target tensor and the predicted tensor.
As defined in https://arxiv.org/abs/1708.02002
Args
y_true: Tensor of target data from the generator with shape (B, N, num_classes).
y_pred: Tensor of predicted data from the network with shape (B, N, num_classes).
Returns
The focal loss of y_pred w.r.t. y_true.
"""
# compute the focal loss
alpha_factor = tf.ones_like(y_true) * alpha
alpha_factor = tf.where(tf.equal(y_true, 1), alpha_factor, 1 - alpha_factor)
focal_weight = tf.where(keras.backend.equal(y_true, 1), 1 - y_pred, y_pred)
focal_weight = alpha_factor * focal_weight ** gamma
cls_loss = focal_weight * tf.keras.backend.binary_crossentropy(y_true, y_pred)
# compute the normalizer: the number of positive anchors
normalizer = tf.cast(tf.shape(y_pred)[1], dtype=tf.float32)
normalizer = tf.maximum(1.0, normalizer)
return tf.reduce_sum(cls_loss) / normalizer
return _focal
def iou():
def _iou(y_true, y_pred):
y_true = tf.maximum(y_true, 0)
pred_left = y_pred[:, :, 0]
pred_top = y_pred[:, :, 1]
pred_right = y_pred[:, :, 2]
pred_bottom = y_pred[:, :, 3]
# (num_pos, )
target_left = y_true[:, :, 0]
target_top = y_true[:, :, 1]
target_right = y_true[:, :, 2]
target_bottom = y_true[:, :, 3]
target_area = (target_left + target_right) * (target_top + target_bottom)
pred_area = (pred_left + pred_right) * (pred_top + pred_bottom)
w_intersect = tf.minimum(pred_left, target_left) + tf.minimum(pred_right, target_right)
h_intersect = tf.minimum(pred_bottom, target_bottom) + tf.minimum(pred_top, target_top)
area_intersect = w_intersect * h_intersect
area_union = target_area + pred_area - area_intersect
# (num_pos, )
iou_loss = -tf.math.log((area_intersect + 1e-7) / (area_union + 1e-7))
# compute the normalizer: the number of positive anchors
normalizer = tf.maximum(1, tf.reduce_prod(tf.shape(y_true)[0:2]))
normalizer = tf.cast(normalizer, dtype=tf.float32)
return tf.reduce_sum(iou_loss) / normalizer
return _iou
def smooth_l1(sigma=3.0):
"""
Create a smooth L1 loss functor.
Args
sigma: This argument defines the point where the loss changes from L2 to L1.
Returns
A functor for computing the smooth L1 loss given target data and predicted data.
"""
sigma_squared = sigma ** 2
def _smooth_l1(y_true, y_pred):
""" Compute the smooth L1 loss of y_pred w.r.t. y_true.
Args
y_true: Tensor from the generator of shape (B, N, 5). The last value for each box is the state of the anchor (ignore, negative, positive).
y_pred: Tensor from the network of shape (B, N, 4).
Returns
The smooth L1 loss of y_pred w.r.t. y_true.
"""
# separate target and state
regression = y_pred
regression_target = y_true[:, :, :-1]
anchor_state = y_true[:, :, -1]
# filter out "ignore" anchors
indices = tf.where(keras.backend.equal(anchor_state, 1))
regression = tf.gather_nd(regression, indices)
regression_target = tf.gather_nd(regression_target, indices)
# compute smooth L1 loss
# f(x) = 0.5 * (sigma * x)^2 if |x| < 1 / sigma / sigma
# |x| - 0.5 / sigma / sigma otherwise
regression_diff = regression - regression_target
regression_diff = keras.backend.abs(regression_diff)
regression_loss = tf.where(
keras.backend.less(regression_diff, 1.0 / sigma_squared),
0.5 * sigma_squared * keras.backend.pow(regression_diff, 2),
regression_diff - 0.5 / sigma_squared
)
# compute the normalizer: the number of positive anchors
normalizer = keras.backend.maximum(1, keras.backend.shape(indices)[0])
normalizer = keras.backend.cast(normalizer, dtype=keras.backend.floatx())
return keras.backend.sum(regression_loss) / normalizer
return _smooth_l1
def focal_with_weight_and_mask(alpha=0.25, gamma=2.0):
"""
Create a functor for computing the focal loss.
Args
alpha: Scale the focal weight with alpha.
gamma: Take the power of the focal weight with gamma.
Returns
A functor that computes the focal loss using the alpha and gamma.
"""
def _focal(inputs):
"""
Compute the focal loss given the target tensor and the predicted tensor.
As defined in https://arxiv.org/abs/1708.02002
Args
y_true: Tensor of target data from the generator with shape (B, N, num_classes).
y_pred: Tensor of predicted data from the network with shape (B, N, num_classes + 2).
Returns
The focal loss of y_pred w.r.t. y_true.
"""
# compute the focal loss
y_true, y_pred, soft_weight, mask = inputs[0][..., :-2], inputs[1], inputs[0][..., -2], inputs[0][..., -1]
alpha_factor = keras.backend.ones_like(y_true) * alpha
alpha_factor = tf.where(keras.backend.equal(y_true, 1), alpha_factor, 1 - alpha_factor)
focal_weight = tf.where(keras.backend.equal(y_true, 1), 1 - y_pred, y_pred)
focal_weight = alpha_factor * focal_weight ** gamma
soft_weight = tf.expand_dims(soft_weight, axis=-1)
cls_loss = focal_weight * soft_weight * tf.keras.backend.binary_crossentropy(y_true, y_pred)
# compute the normalizer: the number of positive locations
num_pos = tf.reduce_sum(mask * soft_weight[..., 0])
normalizer = tf.maximum(1.0, tf.cast(num_pos, dtype=tf.float32))
return tf.reduce_sum(cls_loss) / normalizer
return _focal
def iou_with_weight_and_mask():
def _iou(inputs):
y_true, y_pred, soft_weight, mask = inputs[0][..., :4], inputs[1], inputs[0][..., 4], inputs[0][..., 5]
y_true = tf.maximum(y_true, 0)
pred_left = y_pred[:, :, 0]
pred_top = y_pred[:, :, 1]
pred_right = y_pred[:, :, 2]
pred_bottom = y_pred[:, :, 3]
target_left = y_true[:, :, 0]
target_top = y_true[:, :, 1]
target_right = y_true[:, :, 2]
target_bottom = y_true[:, :, 3]
target_area = (target_left + target_right) * (target_top + target_bottom)
masked_target_area = tf.boolean_mask(target_area, mask)
pred_area = (pred_left + pred_right) * (pred_top + pred_bottom)
masked_pred_area = tf.boolean_mask(pred_area, mask)
w_intersect = tf.minimum(pred_left, target_left) + tf.minimum(pred_right, target_right)
h_intersect = tf.minimum(pred_bottom, target_bottom) + tf.minimum(pred_top, target_top)
area_intersect = w_intersect * h_intersect
masked_area_intersect = tf.boolean_mask(area_intersect, mask)
masked_area_union = masked_target_area + masked_pred_area - masked_area_intersect
masked_soft_weight = tf.boolean_mask(soft_weight, mask)
# (B, N)
masked_iou_loss = -tf.math.log((masked_area_intersect + 1e-7) / (masked_area_union + 1e-7)) * masked_soft_weight
# compute the normalizer: the number of positive locations
num_pos = tf.reduce_sum(mask * soft_weight)
normalizer = keras.backend.maximum(1., num_pos)
return tf.reduce_sum(masked_iou_loss) / normalizer
return _iou