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
|
/*=========================================================================
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkHessianToObjectnessMeasureImageFilter.txx,v $
Language: C++
Date: $Date: 2009-05-11 07:50:26 $
Version: $Revision: 1.5 $
Copyright (c) Insight Software Consortium. All rights reserved.
See ITKCopyright.txt or http://www.itk.org/HTML/Copyright.htm for details.
This software is distributed WITHOUT ANY WARRANTY; without even
the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR
PURPOSE. See the above copyright notices for more information.
=========================================================================*/
#ifndef __itkHessianToObjectnessMeasureImageFilter_txx
#define __itkHessianToObjectnessMeasureImageFilter_txx
#include "itkHessianToObjectnessMeasureImageFilter.h"
#include "itkImageRegionIterator.h"
#include "itkImageRegionConstIterator.h"
#include "itkProgressAccumulator.h"
#include "itkSymmetricEigenAnalysis.h"
#include "vnl/vnl_math.h"
namespace itk
{
/**
* Constructor
*/
template < typename TInputImage, typename TOutputImage >
HessianToObjectnessMeasureImageFilter< TInputImage, TOutputImage>
::HessianToObjectnessMeasureImageFilter()
{
m_Alpha = 0.5;
m_Beta = 0.5;
m_Gamma = 5.0;
m_ScaleObjectnessMeasure = true;
// by default extract bright lines (equivalent to vesselness)
m_ObjectDimension = 1;
m_BrightObject = true;
}
template < typename TInputImage, typename TOutputImage >
void
HessianToObjectnessMeasureImageFilter< TInputImage, TOutputImage>
::BeforeThreadedGenerateData( void )
{
if (m_ObjectDimension >= ImageDimension)
{
itkExceptionMacro( "ObjectDimension must be lower than ImageDimension." );
}
}
template < typename TInputImage, typename TOutputImage >
void
HessianToObjectnessMeasureImageFilter< TInputImage, TOutputImage>
::ThreadedGenerateData(const OutputImageRegionType& outputRegionForThread,
int threadId)
{
typename OutputImageType::Pointer output = this->GetOutput();
typename InputImageType::ConstPointer input = this->GetInput();
// support progress methods/callbacks
ProgressReporter progress(this, threadId, outputRegionForThread.GetNumberOfPixels(), 1000 / this->GetNumberOfThreads() );
// calculator for computation of the eigen values
typedef SymmetricEigenAnalysis< InputPixelType, EigenValueArrayType > CalculatorType;
CalculatorType eigenCalculator( ImageDimension );
// walk the region of eigen values and get the objectness measure
ImageRegionConstIterator<InputImageType> it(input, outputRegionForThread);
ImageRegionIterator<OutputImageType> oit(output, outputRegionForThread );
oit.GoToBegin();
it.GoToBegin();
while (!it.IsAtEnd())
{
// compute eigen values
EigenValueArrayType eigenValues;
eigenCalculator.ComputeEigenValues( it.Get(), eigenValues );
// Sort the eigenvalues by magnitude but retain their sign
EigenValueArrayType sortedEigenValues = eigenValues;
bool done = false;
while (!done)
{
done = true;
for (unsigned int i=0; i<ImageDimension-1; i++)
{
if (vnl_math_abs(sortedEigenValues[i]) > vnl_math_abs(sortedEigenValues[i+1]))
{
std::swap( sortedEigenValues[i], sortedEigenValues[i+1] );
done = false;
}
}
}
// check whether eigenvalues have the right sign
bool signConstraintsSatisfied = true;
for (unsigned int i=m_ObjectDimension; i<ImageDimension; i++)
{
if ((m_BrightObject && sortedEigenValues[i] > 0.0) ||
(!m_BrightObject && sortedEigenValues[i] < 0.0) )
{
signConstraintsSatisfied = false;
break;
}
}
if (!signConstraintsSatisfied)
{
oit.Set(NumericTraits< OutputPixelType >::Zero);
++it;
++oit;
progress.CompletedPixel();
continue;
}
EigenValueArrayType sortedAbsEigenValues;
for (unsigned int i=0; i<ImageDimension; i++)
{
sortedAbsEigenValues[i] = vnl_math_abs(sortedEigenValues[i]);
}
// initialize the objectness measure
double objectnessMeasure = 1.0;
// compute objectness from eigenvalue ratios and second-order structureness
if (m_ObjectDimension < ImageDimension-1)
{
double rA = sortedAbsEigenValues[m_ObjectDimension];
double rADenominatorBase = 1.0;
for (unsigned int j=m_ObjectDimension+1; j<ImageDimension; j++)
{
rADenominatorBase *= sortedAbsEigenValues[j];
}
if (vcl_fabs(rADenominatorBase) > 0.0)
{
if ( vcl_fabs( m_Alpha ) > 0.0 )
{
rA /= vcl_pow(rADenominatorBase, 1.0 / (ImageDimension-m_ObjectDimension-1));
objectnessMeasure *= 1.0 - vcl_exp(- 0.5 * vnl_math_sqr(rA) / vnl_math_sqr(m_Alpha));
}
}
else
{
objectnessMeasure = 0.0;
}
}
if (m_ObjectDimension > 0)
{
double rB = sortedAbsEigenValues[m_ObjectDimension-1];
double rBDenominatorBase = 1.0;
for (unsigned int j=m_ObjectDimension; j<ImageDimension; j++)
{
rBDenominatorBase *= sortedAbsEigenValues[j];
}
if (vcl_fabs(rBDenominatorBase) > 0.0 && vcl_fabs( m_Beta ) > 0.0 )
{
rB /= vcl_pow(rBDenominatorBase, 1.0 / (ImageDimension-m_ObjectDimension));
objectnessMeasure *= vcl_exp(- 0.5 * vnl_math_sqr(rB) / vnl_math_sqr(m_Beta));
}
else
{
objectnessMeasure = 0.0;
}
}
if ( vcl_fabs( m_Gamma ) > 0.0 )
{
double frobeniusNormSquared = 0.0;
for (unsigned int i=0; i<ImageDimension; i++)
{
frobeniusNormSquared += vnl_math_sqr(sortedAbsEigenValues[i]);
}
objectnessMeasure *= 1.0 - vcl_exp(- 0.5 * frobeniusNormSquared / vnl_math_sqr(m_Gamma));
}
// in case, scale by largest absolute eigenvalue
if (m_ScaleObjectnessMeasure)
{
objectnessMeasure *= sortedAbsEigenValues[ImageDimension-1];
}
oit.Set( static_cast< OutputPixelType >(objectnessMeasure));
++it;
++oit;
progress.CompletedPixel();
}
}
template < typename TInputImage, typename TOutputImage >
void
HessianToObjectnessMeasureImageFilter< TInputImage, TOutputImage>
::PrintSelf(std::ostream& os, Indent indent) const
{
Superclass::PrintSelf(os, indent);
os << indent << "Alpha: " << m_Alpha << std::endl;
os << indent << "Beta: " << m_Beta << std::endl;
os << indent << "Gamma: " << m_Gamma << std::endl;
os << indent << "ScaleObjectnessMeasure: " << m_ScaleObjectnessMeasure << std::endl;
os << indent << "ObjectDimension: " << m_ObjectDimension << std::endl;
os << indent << "BrightObject: " << m_BrightObject << std::endl;
}
} // end namespace itk
#endif
|
