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
|
/*=========================================================================
*
* Copyright NumFOCUS
*
* 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
*
* https://www.apache.org/licenses/LICENSE-2.0.txt
*
* 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.
*
*=========================================================================*/
// Disable warning for long symbol names in this file only
#include "itkDiscreteGaussianDerivativeImageFilter.h"
namespace
{
bool
NormalizeSineWave(double frequencyPerImage, unsigned int order, double pixelSpacing = 1.0)
{
// For an image f(x) = sin ( w*x ), where w is a measure of
// frequency, this method verifies that the normalized scale-scale
// is with in reasonable tolerance of the theoretical value.
constexpr unsigned int ImageDimension = 1;
constexpr unsigned int imageSize = 1024;
const double tolerance1 = std::pow(.001, 1.0 / order); // still larger than it should be!
double frequency = frequencyPerImage * 2.0 * itk::Math::pi / (imageSize * pixelSpacing);
// The theoretical maximal value should occur at this sigma
double sigmaMax = std::sqrt(static_cast<double>(order)) / frequency;
// The theoreical maximal value of the derivative, obtained at sigmaMax
double expectedMax = std::pow(static_cast<double>(order), order * 0.5) * std::exp(-0.5 * order);
using ImageType = itk::Image<double, ImageDimension>;
auto image = ImageType::New();
ImageType::SizeType size;
size.Fill(imageSize);
image->SetRegions(ImageType::RegionType(size));
image->Allocate();
ImageType::SpacingType spacing;
spacing.Fill(pixelSpacing);
image->SetSpacing(spacing);
itk::ImageRegionIterator<ImageType> iter(image, image->GetBufferedRegion());
// Create a sine wave image
while (!iter.IsAtEnd())
{
ImageType::PointType p;
image->TransformIndexToPhysicalPoint(iter.GetIndex(), p);
const double x = p[0];
double value = std::sin(x * frequency);
iter.Set(value);
++iter;
}
using GaussianFilterType = itk::DiscreteGaussianDerivativeImageFilter<ImageType, ImageType>;
auto filter = GaussianFilterType::New();
filter->SetInput(image);
filter->SetVariance(itk::Math::sqr(sigmaMax));
filter->SetOrder(order);
filter->SetUseImageSpacing(true);
filter->SetMaximumError(.0005);
filter->SetMaximumKernelWidth(500);
filter->SetNormalizeAcrossScale(true);
ImageType::Pointer outputImage = filter->GetOutput();
outputImage->Update();
// Maximal value of the first derivative
double maxLx = itk::NumericTraits<double>::NonpositiveMin();
itk::ImageRegionConstIterator<ImageType> oiter(outputImage, outputImage->GetBufferedRegion());
while (!oiter.IsAtEnd())
{
maxLx = std::max(maxLx, oiter.Get());
++oiter;
}
// Check if the maximal is obtained with a little bit smaller Gaussian
filter->SetVariance(itk::Math::sqr(sigmaMax * 0.95));
outputImage->Update();
oiter.GoToBegin();
while (!oiter.IsAtEnd())
{
if (maxLx < oiter.Get() && !itk::Math::FloatAlmostEqual(maxLx, oiter.Get(), 10, tolerance1))
{
std::cout.precision(static_cast<int>(itk::Math::abs(std::log10(tolerance1))));
std::cout << "Error at period: " << 1.0 / frequency << std::endl;
std::cout << "Expected maximal value: " << maxLx << ", differs from: " << oiter.Get() << " by more than "
<< tolerance1 << std::endl;
return false;
}
++oiter;
}
// Check if the maximumal value is obtained with a little bit bigger Gaussian
filter->SetVariance(itk::Math::sqr(sigmaMax * 1.05));
outputImage->Update();
oiter.GoToBegin();
while (!oiter.IsAtEnd())
{
if (maxLx < oiter.Get() && !itk::Math::FloatAlmostEqual(maxLx, oiter.Get(), 10, tolerance1))
{
std::cout.precision(static_cast<int>(itk::Math::abs(std::log10(tolerance1))));
std::cout << "Error at period: " << 1.0 / frequency << std::endl;
std::cout << "Expected maximal value: " << maxLx << ", differs from: " << oiter.Get() << " by more than "
<< tolerance1 << std::endl;
return false;
}
++oiter;
}
constexpr double tolerance2 = 0.01;
if (!itk::Math::FloatAlmostEqual(maxLx, expectedMax, 10, tolerance2))
{
std::cout << "Error at frequency: " << frequencyPerImage << std::endl;
std::cout << "Expected maximal value: " << expectedMax << ", differs from: " << maxLx << " by more than "
<< tolerance2 << std::endl;
return false;
}
return true;
}
} // namespace
int
itkDiscreteGaussianDerivativeImageFilterScaleSpaceTest(int, char *[])
{
bool pass = true;
// Testing first order Gaussian
pass &= NormalizeSineWave(16, 1);
pass &= NormalizeSineWave(32, 1);
pass &= NormalizeSineWave(64, 1);
// Testing second order Gaussian
pass &= NormalizeSineWave(16, 2);
pass &= NormalizeSineWave(32, 2);
pass &= NormalizeSineWave(64, 2);
// Testing spacing invariance
pass &= NormalizeSineWave(16, 2, 0.01);
pass &= NormalizeSineWave(16, 2, 100);
if (!pass)
{
std::cout << "Test failed!" << std::endl;
return EXIT_FAILURE;
}
std::cout << "Test finished." << std::endl;
return EXIT_SUCCESS;
}
|
