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
|
/*
//
// Copyright 1997-2012 Torsten Rohlfing
//
// Copyright 2004-2012, 2014 SRI International
//
// This file is part of the Computational Morphometry Toolkit.
//
// http://www.nitrc.org/projects/cmtk/
//
// The Computational Morphometry Toolkit is free software: you can
// redistribute it and/or modify it under the terms of the GNU General Public
// License as published by the Free Software Foundation, either version 3 of
// the License, or (at your option) any later version.
//
// The Computational Morphometry Toolkit is distributed in the hope that it
// will be useful, but WITHOUT ANY WARRANTY; without even the implied
// warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
// GNU General Public License for more details.
//
// You should have received a copy of the GNU General Public License along
// with the Computational Morphometry Toolkit. If not, see
// <http://www.gnu.org/licenses/>.
//
// $Revision: 5436 $
//
// $LastChangedDate: 2018-12-10 19:01:20 -0800 (Mon, 10 Dec 2018) $
//
// $LastChangedBy: torstenrohlfing $
//
*/
#include "cmtkLabelCombinationShapeBasedAveraging.h"
#include <System/cmtkDebugOutput.h>
#include <System/cmtkThreads.h>
#include <Base/cmtkDistanceMap.h>
#include <Base/cmtkUniformDistanceMap.h>
#include <Base/cmtkTypedArray.h>
#include <Base/cmtkTemplateArray.h>
#include <algorithm>
namespace
cmtk
{
/** \addtogroup Segmentation */
//@{
LabelCombinationShapeBasedAveraging::LabelCombinationShapeBasedAveraging( const std::vector<UniformVolume::SmartConstPtr>& labelImages, const Self::LabelIndexType numberOfLabels )
: m_NumberOfLabels( numberOfLabels ),
m_LabelImages( labelImages )
{
if ( ! this->m_NumberOfLabels )
{
this->m_NumberOfLabels = 1;
for ( size_t k = 0; k < this->m_LabelImages.size(); ++k )
{
const Types::DataItemRange range = this->m_LabelImages[k]->GetData()->GetRange();
this->m_NumberOfLabels = std::max( this->m_NumberOfLabels, static_cast<Self::LabelIndexType>( 1 + range.m_UpperBound ) );
}
DebugOutput( 9 ) << "Determined number of labels to be " << this->m_NumberOfLabels << "\n";
}
this->m_NumberOfPixels = this->m_LabelImages[0]->GetNumberOfPixels();
this->m_LabelFlags.resize( this->m_NumberOfLabels, false );
for ( size_t k = 0; k < this->m_LabelImages.size(); ++k )
{
const cmtk::TypedArray& data = *(this->m_LabelImages[k]->GetData());
cmtk::Types::DataItem l;
for ( size_t i = 0; i < this->m_NumberOfPixels; ++i )
{
if ( data.Get( l, i ) )
this->m_LabelFlags[static_cast<unsigned short>( l )] = true;
}
}
}
TypedArray::SmartPtr
LabelCombinationShapeBasedAveraging::GetResult( const bool detectOutliers ) const
{
cmtk::TypedArray::SmartPtr result( cmtk::TypedArray::Create( cmtk::TYPE_USHORT, this->m_NumberOfPixels ) );
result->BlockSet( 0 /*value*/, 0 /*idx*/, this->m_NumberOfPixels /*len*/ );
result->SetDataClass( DATACLASS_LABEL );
std::vector<Self::DistanceMapRealType> totalDistance( this->m_NumberOfPixels, 0.0 );
std::vector<Self::DistanceMapRealType> labelDistanceMap( this->m_NumberOfPixels );
for ( int label = 0; label < this->m_NumberOfLabels; ++label )
{
/// skip labels that are not in any image.
if ( ! this->m_LabelFlags[label] ) continue;
cmtk::DebugOutput( 1 ) << "Processing label #" << label << "\r";
std::fill( labelDistanceMap.begin(), labelDistanceMap.end(), static_cast<Self::DistanceMapRealType>( 0 ) );
if ( detectOutliers )
{
// if this is the first label, write directly to totalDistance, otherwise labelDistanceMap
this->ProcessLabelExcludeOutliers( label, (label == 0) ? totalDistance : labelDistanceMap );
}
else
{
// if this is the first label, write directly to totalDistance, otherwise labelDistanceMap
this->ProcessLabelIncludeOutliers( label, (label == 0) ? totalDistance : labelDistanceMap );
}
// compare sum over all inputs of this label's distance maps pixel by pixel with current total distance map.
// Set result map to this label where it is closer than previous closest label.
if ( label )
{
#pragma omp parallel for
for ( int i = 0; i < static_cast<int>( this->m_NumberOfPixels ); ++i )
{
if ( labelDistanceMap[i] < totalDistance[i] )
{
totalDistance[i] = labelDistanceMap[i];
result->Set( label, i );
}
else
{
if ( !(labelDistanceMap[i] > totalDistance[i]) )
{
result->Set( this->m_NumberOfLabels, i );
}
}
}
}
}
return result;
}
void
LabelCombinationShapeBasedAveraging::ProcessLabelExcludeOutliers
( const Self::LabelIndexType label, std::vector<Self::DistanceMapRealType>& labelDistanceMap ) const
{
const size_t nLabelMaps = this->m_LabelImages.size();
std::vector<cmtk::UniformVolume::SmartConstPtr> signedDistanceMaps( nLabelMaps );
for ( size_t k = 0; k < nLabelMaps; ++k )
{
signedDistanceMaps[k] = cmtk::UniformDistanceMap<Self::DistanceMapRealType>( *(this->m_LabelImages[k]), DistanceMap::VALUE_EXACT + DistanceMap::SIGNED, label ).Get();
}
std::vector<Self::DistanceMapRealType> distances( nLabelMaps );
for ( int i = 0; i < static_cast<int>( this->m_NumberOfPixels ); ++i )
{
for ( size_t k = 0; k < nLabelMaps; ++k )
{
distances[k] = static_cast<Self::DistanceMapRealType>( signedDistanceMaps[k]->GetDataAt( i ) );
}
// sort distance
std::sort( distances.begin(), distances.end() );
// determine 1st and 3rd quartile values
const double Q1 = distances[static_cast<size_t>( 0.25 * nLabelMaps )];
const double Q3 = distances[static_cast<size_t>( 0.75 * nLabelMaps )];
// compute thresholds from quartiles and inter-quartile range
const double lThresh = Q1 - 1.5 * (Q3-Q1);
const double uThresh = Q3 + 1.5 * (Q3-Q1);
for ( size_t k = 0; k < nLabelMaps; ++k )
{
if ( (distances[k] >= lThresh) && (distances[k] <= uThresh) )
labelDistanceMap[i] += distances[k];
}
}
}
void
LabelCombinationShapeBasedAveraging::ProcessLabelIncludeOutliers
( const Self::LabelIndexType label, std::vector<Self::DistanceMapRealType>& labelDistanceMap ) const
{
for ( size_t k = 0; k < this->m_LabelImages.size(); ++k )
{
cmtk::UniformVolume::SmartPtr signedDistanceMap = cmtk::UniformDistanceMap<Self::DistanceMapRealType>( *(this->m_LabelImages[k]), DistanceMap::VALUE_EXACT + DistanceMap::SIGNED, label ).Get();
const Self::DistanceMapRealType* signedDistancePtr = static_cast<const Self::DistanceMapRealType*>( signedDistanceMap->GetData()->GetDataPtr() );
#pragma omp parallel for
for ( int i = 0; i < static_cast<int>( this->m_NumberOfPixels ); ++i )
{
labelDistanceMap[i] += signedDistancePtr[i];
}
}
}
} // namespace cmtk
|
