This repository has been archived by the owner on May 23, 2022. It is now read-only.
-
Notifications
You must be signed in to change notification settings - Fork 7
/
test_feature_distribution_matching.py
313 lines (246 loc) · 12.6 KB
/
test_feature_distribution_matching.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
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
# pylint: disable=missing-module-docstring (C0114)
# pylint: disable=missing-function-docstring (C0116)
# pylint: disable=protected-access (W0212)
import inspect
from typing import Callable, Tuple
from unittest.mock import MagicMock, Mock, patch
import numpy as np
import pytest
from matching import ChannelsType
from matching.operations import \
FeatureDistributionMatching as FeatureDistMatching
from tests import (CHANNEL_RANGES_DEFAULT, CHANNELS_DEFAULT, MUNICH_1_PATH,
MUNICH_2_PATH)
from utils.cs_conversion import ChannelRange
from utils.image_io import read_image
ImageGenType = Callable[[Tuple[int, int, int], int], np.ndarray]
TEST_IMAGE_423 = np.array([[[0.09634835, 0.67985358, 0.71879272],
[0.2746647, 0.55942712, 0.17269985]],
[[0.60345517, 0.70931529, 0.14624073],
[0.23241476, 0., 0.91803395]], # 0. entry
[[0.58758528, 0.66624122, 0.55438404],
[1., 0.26015386, 0.28256821]], # 1. entry
[[0.87368081, 0.85794979, 0.11262025],
[0.70622847, 0.9368422, 0.39187311]]])
TEST_IMAGE_421 = TEST_IMAGE_423[:, :, 0:1]
TEST_IMAGE_243 = np.array([[[0.96883339, 0.6280047, 0.55266079],
[0.64656832, 0.96452021, 0.01043656],
[1., 0.75418668, 0.77878355], # 1. entry
[0.80548047, 0.73128374, 0.72436705]],
[[0.81233289, 0.67804285, 0.10213132],
[0.3819547, 0.67940864, 0.8103251],
[0.30974552, 0.49638342, 0.], # 0. entry
[0.11005092, 0.48727506, 0.31166669]]])
TEST_IMAGE_241 = TEST_IMAGE_243[:, :, 0:1]
def test_channel_ranges() -> None:
with pytest.raises(TypeError):
FeatureDistMatching(CHANNELS_DEFAULT, check_input=True,
channel_ranges=[1, 2, 3]) # type: ignore
with pytest.raises(TypeError):
FeatureDistMatching(CHANNELS_DEFAULT, check_input=True,
channel_ranges=(
ChannelRange(0.1, 1.7), 1, # type: ignore
ChannelRange(0.1, 1.7)))
@pytest.mark.parametrize('test_image, des_shape',
[(TEST_IMAGE_423, (8, 3)),
(TEST_IMAGE_421, (8, 1))
])
def test_get_feature_matrix(test_image: np.array,
des_shape: Tuple[int, int]) -> None:
result = FeatureDistMatching._get_feature_matrix(test_image)
np.testing.assert_array_equal(result, test_image.reshape(des_shape))
def test_center_image() -> None:
test_image = TEST_IMAGE_423
test_image_mat = FeatureDistMatching._get_feature_matrix(test_image)
test_image_mean = np.mean(test_image_mat, axis=0)
result_image_mat, result_mean = FeatureDistMatching._center_image(
test_image_mat)
np.testing.assert_array_equal(result_mean, test_image_mean)
np.testing.assert_almost_equal(np.mean(result_image_mat, axis=0),
np.zeros(3))
def test_whitening() -> None:
test_image = TEST_IMAGE_423
test_image_mat = FeatureDistMatching._get_feature_matrix(test_image)
FeatureDistMatching._center_image(test_image_mat)
result = FeatureDistMatching._whitening(test_image_mat)
np.testing.assert_almost_equal(np.cov(result, rowvar=False),
np.identity(3))
def test_whitening_2d() -> None:
test_image = TEST_IMAGE_421
test_image_mat = FeatureDistMatching._get_feature_matrix(test_image)
FeatureDistMatching._center_image(test_image_mat)
result = FeatureDistMatching._whitening(test_image_mat)
np.testing.assert_almost_equal(np.var(result), 1.)
def test_covariance_transformation() -> None:
test_image = TEST_IMAGE_423
test_image_mat = FeatureDistMatching._get_feature_matrix(test_image)
FeatureDistMatching._center_image(test_image_mat)
test_image_white = FeatureDistMatching._whitening(test_image_mat)
feature_mat_ref = FeatureDistMatching._get_feature_matrix(TEST_IMAGE_243)
result = FeatureDistMatching._covariance_transformation(test_image_white,
feature_mat_ref)
np.testing.assert_almost_equal(np.cov(result, rowvar=False),
np.cov(feature_mat_ref, rowvar=False))
def test_covariance_transformation_2d() -> None:
test_image = TEST_IMAGE_421
test_image_mat = FeatureDistMatching._get_feature_matrix(test_image)
FeatureDistMatching._center_image(test_image_mat)
test_image_white = FeatureDistMatching._whitening(test_image_mat)
feature_mat_ref = FeatureDistMatching._get_feature_matrix(
TEST_IMAGE_243[:, :, 0])
result = FeatureDistMatching._covariance_transformation(test_image_white,
feature_mat_ref)
np.testing.assert_almost_equal(np.var(result),
np.var(feature_mat_ref))
@pytest.fixture(name='feature_dist_matching')
def fixture_feature_dist_matching() -> FeatureDistMatching:
return FeatureDistMatching(CHANNELS_DEFAULT,
CHANNEL_RANGES_DEFAULT,
check_input=True)
def test_design() -> None:
assert inspect.isabstract(FeatureDistMatching) is False
assert len(FeatureDistMatching.__mro__) == 4
@pytest.mark.parametrize('channel_range_max', [1.0, 255.])
def test_matching(channel_range_max: float) -> None:
source = TEST_IMAGE_423
reference = TEST_IMAGE_243
source[:, :, 1] *= channel_range_max
reference[:, :, 1] *= channel_range_max
result = FeatureDistMatching._matching(source, reference)
feature_mat_result = FeatureDistMatching._get_feature_matrix(result)
feature_mat_ref = FeatureDistMatching._get_feature_matrix(reference)
result_mean = np.mean(feature_mat_result, axis=0)
reference_mean = np.mean(feature_mat_ref, axis=0)
np.testing.assert_almost_equal(result_mean, reference_mean)
feature_mat_result -= result_mean
feature_mat_ref -= reference_mean
np.testing.assert_almost_equal(np.cov(feature_mat_result, rowvar=False),
np.cov(feature_mat_ref, rowvar=False))
@pytest.mark.parametrize('channel_range_max', [1.0, 255.])
def test_matching_2d(channel_range_max: float) -> None:
source = TEST_IMAGE_421 * channel_range_max
reference = TEST_IMAGE_421 * channel_range_max
result = FeatureDistMatching._matching(source, reference)
feature_mat_result = FeatureDistMatching._get_feature_matrix(result)
feature_mat_ref = FeatureDistMatching._get_feature_matrix(reference)
result_mean = np.mean(feature_mat_result, axis=0)
reference_mean = np.mean(feature_mat_ref, axis=0)
np.testing.assert_almost_equal(result_mean, reference_mean)
feature_mat_result -= result_mean
feature_mat_ref -= reference_mean
np.testing.assert_almost_equal(np.var(feature_mat_result),
np.var(feature_mat_ref))
@pytest.mark.parametrize('source, reference',
[(TEST_IMAGE_423, TEST_IMAGE_243),
(TEST_IMAGE_421, TEST_IMAGE_241)])
def test_apply(feature_dist_matching: FeatureDistMatching,
source: np.array,
reference: np.array) -> None:
feature_dist_matching.channels = CHANNELS_DEFAULT[0:source.shape[-1]]
feature_dist_matching.channel_ranges = CHANNEL_RANGES_DEFAULT[
0:source.shape[-1]]
source_copy = np.copy(source)
reference_copy = np.copy(reference)
result = feature_dist_matching._apply(source, reference)
# Check modification of input:
np.testing.assert_array_equal(source, source_copy)
np.testing.assert_array_equal(reference, reference_copy)
# Check result:
assert result.shape == source.shape
assert np.max(result) <= 1.
assert 0.1 < np.mean(result) < 0.9
assert np.min(result) >= 0.
assert result.dtype == np.float32
@pytest.mark.parametrize('channels',
[(0,), (1,), (2,), (0, 1), (0, 2), (1, 2), (0, 1, 2)])
def test_apply_channels(channels: ChannelsType,
feature_dist_matching: FeatureDistMatching) -> None:
source = TEST_IMAGE_423.astype(np.float32)
reference = TEST_IMAGE_243.astype(np.float32)
feature_dist_matching.channels = channels
result = feature_dist_matching._apply(source, reference)
for untouched_channel in {0, 1, 2} - set(channels):
np.testing.assert_array_equal(source[:, :, untouched_channel],
result[:, :, untouched_channel])
def test_apply_channel_range(
feature_dist_matching: FeatureDistMatching) -> None:
source = TEST_IMAGE_423
reference = TEST_IMAGE_243
source[:, :, 1] *= 255.
source[:, :, 2] *= 254.
source[:, :, 2] -= 127.
reference[:, :, 1] *= 255.
reference[:, :, 2] *= 254.
reference[:, :, 2] -= 127.
feature_dist_matching.channel_ranges = tuple([ChannelRange(0., 1.),
ChannelRange(0., 255.),
ChannelRange(-127.0, 127.0)])
result = feature_dist_matching._apply(source, reference)
np.testing.assert_array_equal(source, TEST_IMAGE_423)
np.testing.assert_array_equal(reference, TEST_IMAGE_243)
assert result.shape == source.shape
assert result.dtype == np.float32
assert np.max(result[:, :, 0]) <= 1.
assert 0.1 < np.mean(result[:, :, 0]) < 0.9
assert np.max(result[:, :, 0]) >= 0.
assert np.max(result[:, :, 1]) <= 255.
assert 5. < np.mean(result[:, :, 1]) < 250.
assert np.min(result[:, :, 1]) >= 0.
assert np.max(result[:, :, 2]) <= 127.
assert -120. < np.mean(result[:, :, 2]) < 120.
assert np.min(result[:, :, 2]) >= -127.
@patch.object(FeatureDistMatching,
FeatureDistMatching._get_feature_matrix.__name__)
@patch.object(FeatureDistMatching,
FeatureDistMatching._center_image.__name__,
return_value=(np.ones((2, 3)), np.ones((2, 3))))
@patch.object(FeatureDistMatching, FeatureDistMatching._whitening.__name__)
@patch.object(FeatureDistMatching,
FeatureDistMatching._covariance_transformation.__name__)
def test_matching_order(covariance_transformation: MagicMock,
whitening: MagicMock,
center_image: MagicMock,
get_feature_matrix: MagicMock,
feature_dist_matching: FeatureDistMatching) -> None:
mock = Mock()
mock_get_feature_matrix = 'mock_get_feature_matrix'
mock_center_image = 'mock_center_image'
mock_whitening = 'mock_whitening'
mock_covariance_transformation = 'mock_covariance_transformation'
mock.attach_mock(get_feature_matrix, mock_get_feature_matrix)
mock.attach_mock(center_image, mock_center_image)
mock.attach_mock(whitening, mock_whitening)
mock.attach_mock(covariance_transformation, mock_covariance_transformation)
source = TEST_IMAGE_423
reference = TEST_IMAGE_243
feature_dist_matching._matching(source, reference)
# check the call order
expected_calls = [mock_get_feature_matrix,
mock_get_feature_matrix,
mock_center_image,
mock_center_image,
mock_whitening,
mock_covariance_transformation,
mock_covariance_transformation + '().__iadd__',
mock_covariance_transformation + '().__iadd__().reshape']
calls = [call[0] for call in mock.mock_calls]
assert calls == expected_calls
@pytest.mark.parametrize('source_path, reference_path',
[(MUNICH_1_PATH, MUNICH_2_PATH),
(MUNICH_2_PATH, MUNICH_1_PATH),
(MUNICH_1_PATH, MUNICH_1_PATH)])
def test_apply_images(source_path: str, reference_path: str,
feature_dist_matching: FeatureDistMatching) -> \
None:
source = read_image(source_path)
reference = read_image(reference_path)
result = feature_dist_matching._apply(source, reference)
assert result.shape == source.shape
assert np.max(result) <= 1.
assert 0.1 < np.mean(result) < 0.9
assert np.min(result) >= 0.
assert result.dtype == np.float32
with np.testing.assert_raises(AssertionError):
np.testing.assert_array_equal(result, source)
with np.testing.assert_raises(AssertionError):
np.testing.assert_array_equal(result, reference)