forked from InsightSoftwareConsortium/ITK
-
Notifications
You must be signed in to change notification settings - Fork 0
/
itkShapePriorMAPCostFunction.h
167 lines (127 loc) · 6.4 KB
/
itkShapePriorMAPCostFunction.h
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
/*=========================================================================
Program: Insight Segmentation & Registration Toolkit
Module: itkShapePriorMAPCostFunction.h
Language: C++
Date: $Date$
Version: $Revision$
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 __itkShapePriorMAPCostFunction_h
#define __itkShapePriorMAPCostFunction_h
#include "itkShapePriorMAPCostFunctionBase.h"
#include "itkGaussianKernelFunction.h"
namespace itk
{
/** \class ShapePriorMAPCostFunction
* \brief Represents the maximum aprior (MAP) cost function used
* ShapePriorSegmentationLevelSetImageFilter to estimate the shape paramaeters.
*
* This class follows the shape and pose parameters estimation developed in [1].
* Note that this class returns the negative log of the MAP function.
* Using the negative function make this cost function compatible
* with generic optimizers which seeks the minimum of a cost function.
*
* This class has two template parameters, the feature image type representing the
* edge potential map and the pixel type used to
* represent the output level set in the ShapePriorSegmentationLevelSetImageFilter.
*
* \sa ShapePriorSegmentationLevelSetImageFilter
*
* \par REFERENCES
* \par
* [1] Leventon, M.E. et al. "Statistical Shape Influence in Geodesic Active Contours", CVPR 2000.
*
*
* \ingroup Numerics Optimizers
*/
template <class TFeatureImage, class TOutputPixel>
class ITK_EXPORT ShapePriorMAPCostFunction :
public ShapePriorMAPCostFunctionBase< TFeatureImage, TOutputPixel>
{
public:
/** Standard class typedefs. */
typedef ShapePriorMAPCostFunction Self;
typedef ShapePriorMAPCostFunctionBase<TFeatureImage,TOutputPixel> Superclass;
typedef SmartPointer<Self> Pointer;
typedef SmartPointer<const Self> ConstPointer;
/** Method for creation through the object factory. */
itkNewMacro(Self);
/** Run-time type information (and related methods). */
itkTypeMacro( ShapePriorMAPCostFunction, ShapePriorMAPCostFunctionBase );
/** ParametersType typedef.
* It defines a position in the optimization search space. */
typedef typename Superclass::ParametersType ParametersType;
/** Type of the feature image representing the edge potential map. */
typedef typename Superclass::FeatureImageType FeatureImageType;
typedef typename Superclass::FeatureImagePointer FeatureImagePointer;
/** Type of the return measure value. */
typedef typename Superclass::MeasureType MeasureType;
/** Dimension constant. */
itkStaticConstMacro( ImageDimension, unsigned int, TFeatureImage::ImageDimension);
/** Type of pixel used to represent the level set. */
typedef typename Superclass::PixelType PixelType;
/** Type of node used to represent the active region around the zero set. */
typedef typename Superclass::NodeType NodeType;
/** Type of container used to store the level set nodes. */
typedef typename Superclass::NodeContainerType NodeContainerType;
/** Type of the shape signed distance function. */
typedef typename Superclass::ShapeFunctionType ShapeFunctionType;
/** Type of the array for storing shape parameter mean and standard deivation. */
typedef Array<double> ArrayType;
/** Set/Get the array of shape parameters mean. */
itkSetMacro( ShapeParameterMeans, ArrayType );
itkGetMacro( ShapeParameterMeans, ArrayType );
/** Set/Get the array of shape parameters standard deviation. */
itkSetMacro( ShapeParameterStandardDeviations, ArrayType );
itkGetMacro( ShapeParameterStandardDeviations, ArrayType );
/** Set/Get the weights for each term. Default is a vector of all ones.
* The weights are applied to terms in the following order:
* LogInsideTerm, LogGradientTerm, LogShapePriorTerm and LogPosePriorTerm.*/
typedef FixedArray<double,4> WeightsType;
itkSetMacro( Weights, WeightsType );
itkGetConstReferenceMacro( Weights, WeightsType );
/** Compute the inside term component of the MAP cost function.
* In particular, the method sums the number of pixels inside
* the current contour (defined by nodes of the active region
* that are less than zero) which are outside the shape
* specified by the input parameters. */
virtual MeasureType ComputeLogInsideTerm( const ParametersType & parameters ) const;
/** Compute the gradient term component of the MAP cost function.
* In particular, this method assume that ( 1 - FeatureImage ) approximates
* a Gaussian (zero mean, unit variance) algon the normal of the evolving contour.
* The gradient term is then given by a Laplacian of the goodness of fit of
* the Gaussian. */
virtual MeasureType ComputeLogGradientTerm( const ParametersType & parameters ) const;
/** Compute the shape prior component of the MAP cost function.
* In particular, the method assumes that the shape parameters comes from
* independent Gaussian distributions defined by the ShapeParameterMeans
* and ShapeParameterVariances array. */
virtual MeasureType ComputeLogShapePriorTerm( const ParametersType & parameters ) const;
/** Compute the pose prior component of the MAP cost function.
* In particular, this method assumes that the pose parameters are
* uniformly distributed and returns a constant of zero. */
virtual MeasureType ComputeLogPosePriorTerm( const ParametersType & parameters ) const;
/** Initialize the cost function by making sure that all the components
* are present. */
virtual void Initialize(void) throw ( ExceptionObject );
protected:
ShapePriorMAPCostFunction();
virtual ~ShapePriorMAPCostFunction() {};
void PrintSelf(std::ostream& os, Indent indent) const;
private:
ShapePriorMAPCostFunction(const Self&); //purposely not implemented
void operator=(const Self&); //purposely not implemented
ArrayType m_ShapeParameterMeans;
ArrayType m_ShapeParameterStandardDeviations;
WeightsType m_Weights;
typename GaussianKernelFunction::Pointer m_GaussianFunction;
};
} // end namespace itk
#ifndef ITK_MANUAL_INSTANTIATION
#include "itkShapePriorMAPCostFunction.txx"
#endif
#endif