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
|
# Copyright (c) DataLab Platform Developers, BSD 3-Clause license, see LICENSE file.
"""Unit tests for image convolution/deconvolution features."""
from __future__ import annotations
from contextlib import contextmanager
import numpy as np
import pytest
import scipy.signal as sps
import sigima.proc.image
from sigima.config import options as sigima_options
from sigima.objects import create_image, create_image_from_param
from sigima.objects.image import Gauss2DParam, ImageObj, Zero2DParam
from sigima.tests import guiutils
from sigima.tests.helpers import check_array_result
from sigima.tools.image import deconvolve
@contextmanager
def disable_kernel_normalization():
"""Context manager to temporarily disable kernel normalization."""
# Save current values
original_auto_normalize = sigima_options.auto_normalize_kernel.get()
# Disable for test
sigima_options.auto_normalize_kernel.set(False)
try:
yield
finally:
# Restore original values
sigima_options.auto_normalize_kernel.set(original_auto_normalize)
def _generate_rectangle_image(title: str = "Rectangle", size: int = 32) -> ImageObj:
"""Generate a test square image with a rectangle in the center."""
data = np.zeros((size, size), dtype=np.float64)
data[size // 5 : 2 * size // 5, size // 7 : 5 * size // 7] = 1.0
img = create_image(title, data)
return img
def _generate_image(size: int = 32) -> ImageObj:
"""Generate a test square image.
Args:
size: The dimension of the square image to generate.
Returns:
An image object.
"""
# Gaussian image.
gauss_img = create_image_from_param(Gauss2DParam.create(height=size, width=size))
return gauss_img
def _generate_gaussian_kernel(size: int = 32, sigma: float = 1.0) -> ImageObj:
"""Generate a Gaussian kernel image.
Args:
size: The dimension of the square kernel to generate.
sigma: The standard deviation of the Gaussian.
Returns:
An image object.
"""
kernel = create_image_from_param(
Gauss2DParam.create(height=size, width=size, sigma=sigma)
)
return kernel
def _generate_identity_kernel(size: int = 7, ignore_odd: bool = False) -> ImageObj:
"""Generate an identity kernel image.
Args:
size: The dimension of the square kernel to generate (must be odd).
ignore_odd: If True, do not check for odd size.
Returns:
An image object.
"""
if not ignore_odd:
assert size % 2 == 1, "Identity kernel size must be odd."
kernel = create_image_from_param(Zero2DParam.create(height=size, width=size))
assert kernel.data is not None
kernel.data[size // 2, size // 2] = 1.0
return kernel
def __convolve_image(kernel: ImageObj, size: int = 32) -> tuple[ImageObj, ImageObj]:
"""Generate a test image and its convolution with a Gaussian kernel.
Returns:
A tuple (source image, kernel, convolved image).
"""
original = _generate_rectangle_image(size=size)
assert original.data is not None
convolved = sigima.proc.image.convolution(original, kernel)
assert convolved.data is not None
return original, convolved
@pytest.mark.validation
def test_image_convolution(size: int = 32) -> None:
"""Validation test for the image convolution processing.
Note: This test disables kernel normalization to compare against raw scipy results.
"""
with disable_kernel_normalization():
# Test with a Gaussian kernel.
kernel = _generate_gaussian_kernel(size=size, sigma=1.0)
original, convolved = __convolve_image(kernel, size=size)
exp = sps.convolve(original.data, kernel.data, mode="same", method="auto")
check_array_result("Convolution", convolved.data, exp)
guiutils.view_images_side_by_side_if_gui(
[original, kernel, convolved],
["Original", "Kernel", "Convolved"],
title="Image Convolution Test: Gaussian Kernel",
)
# Test with an identity kernel.
kernel = _generate_identity_kernel(size=7)
original, convolved = __convolve_image(kernel, size=size)
# check_array_result("Convolution identity", convolved.data, original.data)
guiutils.view_images_side_by_side_if_gui(
[original, kernel, convolved],
["Original", "Kernel", "Convolved"],
title="Image Convolution Test: Identity Kernel",
)
# Test with a null kernel.
kernel.data = np.array([])
with pytest.raises(ValueError, match="Convolution kernel cannot be null."):
sigima.proc.image.convolution(original, kernel)
@pytest.mark.validation
def test_image_deconvolution(size: int = 32) -> None:
"""Validation test for image deconvolution.
Note: This test disables kernel normalization to compare against expected results.
"""
with disable_kernel_normalization():
# Test with an identity kernel.
kernel = _generate_identity_kernel(size=31)
original, convolved = __convolve_image(kernel, size=size)
deconvolved = sigima.proc.image.deconvolution(convolved, kernel)
guiutils.view_images_side_by_side_if_gui(
[original, kernel, convolved, deconvolved],
["Original", "Kernel", "Convolved", "Deconvolved"],
title="Image Deconvolution Test: Identity Kernel",
)
check_array_result("Deconvolution identity", deconvolved.data, original.data)
# Test with a Gaussian kernel.
kernel = _generate_gaussian_kernel(size=31, sigma=1.0)
original, convolved = __convolve_image(kernel, size=64)
deconvolved = sigima.proc.image.deconvolution(convolved, kernel)
# check_array_result("Deconvolution Gaussian", deconvolved.data, original.data)
guiutils.view_images_side_by_side_if_gui(
[original, kernel, convolved, deconvolved],
["Original", "Kernel", "Convolved", "Deconvolved"],
title="Image Deconvolution Test: Gaussian Kernel",
)
# Test with a non-odd-sized kernel.
kernel = _generate_identity_kernel(size=8, ignore_odd=True)
# Check that a warning is raised for non-odd-sized kernel.
with pytest.warns(
UserWarning, match=r"Deconvolution kernel has even dimension\(s\).*"
):
deconvolved = sigima.proc.image.deconvolution(convolved, kernel)
def test_tools_image_deconvolve_null_kernel() -> None:
"""Test deconvolution with a null kernel."""
size = 32
src = _generate_image(size)
assert src.data is not None
kernel = create_image_from_param(Zero2DParam.create(height=size, width=size))
assert kernel.data is not None
with pytest.raises(ValueError, match="Deconvolution kernel cannot be null."):
deconvolve(src.data, kernel.data)
if __name__ == "__main__":
guiutils.enable_gui()
test_image_convolution()
test_image_deconvolution()
test_tools_image_deconvolve_null_kernel()
|
