-
Notifications
You must be signed in to change notification settings - Fork 610
/
utils.py
151 lines (124 loc) · 4.82 KB
/
utils.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
# Copyright 2019 The TensorFlow Authors. All Rights Reserved.
#
# 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.
# ==============================================================================
"""Image util ops."""
import tensorflow as tf
def get_ndims(image):
return image.get_shape().ndims or tf.rank(image)
def to_4D_image(image):
"""Convert 2/3/4D image to 4D image.
Args:
image: 2/3/4D `Tensor`.
Returns:
4D `Tensor` with the same type.
"""
with tf.control_dependencies(
[
tf.debugging.assert_rank_in(
image, [2, 3, 4], message="`image` must be 2/3/4D tensor"
)
]
):
ndims = image.get_shape().ndims
if ndims is None:
return _dynamic_to_4D_image(image)
elif ndims == 2:
return image[None, :, :, None]
elif ndims == 3:
return image[None, :, :, :]
else:
return image
def _dynamic_to_4D_image(image):
shape = tf.shape(image)
original_rank = tf.rank(image)
# 4D image => [N, H, W, C] or [N, C, H, W]
# 3D image => [1, H, W, C] or [1, C, H, W]
# 2D image => [1, H, W, 1]
left_pad = tf.cast(tf.less_equal(original_rank, 3), dtype=tf.int32)
right_pad = tf.cast(tf.equal(original_rank, 2), dtype=tf.int32)
new_shape = tf.concat(
[
tf.ones(shape=left_pad, dtype=tf.int32),
shape,
tf.ones(shape=right_pad, dtype=tf.int32),
],
axis=0,
)
return tf.reshape(image, new_shape)
def from_4D_image(image, ndims):
"""Convert back to an image with `ndims` rank.
Args:
image: 4D `Tensor`.
ndims: The original rank of the image.
Returns:
`ndims`-D `Tensor` with the same type.
"""
with tf.control_dependencies(
[tf.debugging.assert_rank(image, 4, message="`image` must be 4D tensor")]
):
if isinstance(ndims, tf.Tensor):
return _dynamic_from_4D_image(image, ndims)
elif ndims == 2:
return tf.squeeze(image, [0, 3])
elif ndims == 3:
return tf.squeeze(image, [0])
else:
return image
def _dynamic_from_4D_image(image, original_rank):
shape = tf.shape(image)
# 4D image <= [N, H, W, C] or [N, C, H, W]
# 3D image <= [1, H, W, C] or [1, C, H, W]
# 2D image <= [1, H, W, 1]
begin = tf.cast(tf.less_equal(original_rank, 3), dtype=tf.int32)
end = 4 - tf.cast(tf.equal(original_rank, 2), dtype=tf.int32)
new_shape = shape[begin:end]
return tf.reshape(image, new_shape)
def wrap(image):
"""Returns `image` with an extra channel set to all 1s."""
shape = tf.shape(image)
extended_channel = tf.ones([shape[0], shape[1], 1], image.dtype)
extended = tf.concat([image, extended_channel], 2)
return extended
def unwrap(image, replace):
"""Unwraps an image produced by wrap.
Where there is a 0 in the last channel for every spatial position,
the rest of the three channels in that spatial dimension are grayed
(set to 128). Operations like translate and shear on a wrapped
Tensor will leave 0s in empty locations. Some transformations look
at the intensity of values to do preprocessing, and we want these
empty pixels to assume the 'average' value, rather than pure black.
Args:
image: A 3D image `Tensor` with 4 channels.
replace: A one or three value 1D `Tensor` to fill empty pixels.
Returns:
image: A 3D image `Tensor` with 3 channels.
"""
image_shape = tf.shape(image)
# Flatten the spatial dimensions.
flattened_image = tf.reshape(image, [-1, image_shape[2]])
# Find all pixels where the last channel is zero.
alpha_channel = flattened_image[:, 3]
replace = tf.constant(replace, tf.uint8)
if tf.rank(replace) == 0:
replace = tf.expand_dims(replace, 0)
replace = tf.concat([replace, replace, replace], 0)
replace = tf.concat([replace, tf.ones([1], dtype=image.dtype)], 0)
# Where they are zero, fill them in with 'replace'.
cond = tf.equal(alpha_channel, 1)
cond = tf.expand_dims(cond, 1)
cond = tf.concat([cond, cond, cond, cond], 1)
flattened_image = tf.where(cond, flattened_image, replace)
image = tf.reshape(flattened_image, image_shape)
image = tf.slice(image, [0, 0, 0], [image_shape[0], image_shape[1], 3])
return image