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
|
/*=========================================================================
*
* 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.
*
*=========================================================================*/
#include "itkHessianRecursiveGaussianImageFilter.h"
// This test creates an image varying as a 1D Gaussian in the X direction
// for different values of sigma, and checks the scale-space response of
// the xx component of the Hessian at the center of the Gaussian.
// If NormalizeAcrossScale works correctly, the filter should yield the
// same Hxx across different scales.
int
itkHessianRecursiveGaussianFilterScaleSpaceTest(int, char *[])
{
constexpr unsigned int Dimension = 3;
using PixelType = double;
using ImageType = itk::Image<PixelType, Dimension>;
using IndexType = itk::Index<Dimension>;
using SizeType = itk::Size<Dimension>;
using RegionType = itk::ImageRegion<Dimension>;
using PointType = ImageType::PointType;
using SpacingType = ImageType::SpacingType;
auto inputImage = ImageType::New();
SizeType size;
size.Fill(21);
size[0] = 401;
IndexType start;
start.Fill(0);
RegionType region;
region.SetIndex(start);
region.SetSize(size);
PointType origin;
origin.Fill(-1.25);
origin[0] = -20.0;
SpacingType spacing;
spacing.Fill(0.1);
inputImage->SetOrigin(origin);
inputImage->SetSpacing(spacing);
inputImage->SetRegions(region);
inputImage->Allocate();
using IteratorType = itk::ImageRegionIteratorWithIndex<ImageType>;
constexpr unsigned int numberOfScales = 4;
double scales[numberOfScales];
scales[0] = 1.0;
scales[1] = 2.0;
scales[2] = 3.0;
scales[3] = 5.0;
// changing the size of the object with the size of the
// gaussian should produce the same results
for (double objectSize : scales)
{
IteratorType it(inputImage, inputImage->GetRequestedRegion());
PointType point;
// Fill the image with a 1D Gaussian along X with sigma equal to the current scale
// The Gaussian is not normalized, since it should have the same peak value across
// scales, only sigma should change
while (!it.IsAtEnd())
{
inputImage->TransformIndexToPhysicalPoint(it.GetIndex(), point);
double value = std::exp(-point[0] * point[0] / (2.0 * objectSize * objectSize));
it.Set(value);
++it;
}
// Compute the hessian using NormalizeAcrossScale true
using FilterType = itk::HessianRecursiveGaussianImageFilter<ImageType>;
using HessianImageType = FilterType::OutputImageType;
auto filter = FilterType::New();
filter->SetInput(inputImage);
filter->SetSigma(objectSize);
filter->SetNormalizeAcrossScale(true);
filter->Update();
HessianImageType::Pointer outputImage = filter->GetOutput();
// Get the value at the center of the image, the location of the peak of the Gaussian
PointType center;
center.Fill(0.0);
IndexType centerIndex = outputImage->TransformPhysicalPointToIndex(center);
// Irrespective of the scale, the Hxx component should be the same
double centerHxx = outputImage->GetPixel(centerIndex)[0];
if (centerHxx > -0.3546 || centerHxx < -0.3547)
{
std::cout << "center Hessian: " << outputImage->GetPixel(centerIndex) << std::endl;
return EXIT_FAILURE;
}
}
// maintaining the size of the object and gaussian, in physical
// size, should maintain the value, while the size of the image changes.
for (double scale : scales)
{
IteratorType it(inputImage, inputImage->GetRequestedRegion());
PointType point;
double objectSize = 5.0;
spacing.Fill(scale / 5.0);
inputImage->SetSpacing(spacing);
// Fill the image with a 1D Gaussian along X with sigma equal to
// the object size.
// The Gaussian is not normalized, since it should have the same peak value across
// scales, only sigma should change
while (!it.IsAtEnd())
{
inputImage->TransformIndexToPhysicalPoint(it.GetIndex(), point);
double value = std::exp(-point[0] * point[0] / (2.0 * objectSize * objectSize));
it.Set(value);
++it;
}
// Compute the hessian using NormalizeAcrossScale true
using FilterType = itk::HessianRecursiveGaussianImageFilter<ImageType>;
using HessianImageType = FilterType::OutputImageType;
auto filter = FilterType::New();
filter->SetInput(inputImage);
filter->SetSigma(objectSize);
filter->SetNormalizeAcrossScale(true);
filter->Update();
HessianImageType::Pointer outputImage = filter->GetOutput();
// Get the value at the center of the image, the location of the peak of the Gaussian
PointType center;
center.Fill(0.0);
IndexType centerIndex = outputImage->TransformPhysicalPointToIndex(center);
// Irrespective of the scale, the Hxx component should be the same
double centerHxx = outputImage->GetPixel(centerIndex)[0];
if (centerHxx > -0.354 || centerHxx < -0.355)
{
std::cout << "center Hessian: " << outputImage->GetPixel(centerIndex) << std::endl;
return EXIT_FAILURE;
}
}
return EXIT_SUCCESS;
}
|
