From 48532460367ab7b72ae149c7b93b0678cc469cf1 Mon Sep 17 00:00:00 2001
From: Romain Garrigues <romain.garrigues@c-s.fr>
Date: Wed, 28 Jun 2006 15:35:15 +0000
Subject: [PATCH] =?UTF-8?q?Mise=20=C3=A0=20jour=20ITK=20version=202.8.1?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
---
Utilities/ITK/Code/Numerics/CMakeLists.txt | 9 +-
.../ITK/Code/Numerics/FEM/CMakeLists.txt | 7 +-
.../itkFEMElement2DC0LinearQuadrilateral.cxx | 6 +-
.../Numerics/FEM/itkFEMElement2DC1Beam.cxx | 8 +-
.../FEM/itkFEMElement3DC0LinearHexahedron.cxx | 6 +-
.../itkFEMElement3DC0LinearTetrahedron.cxx | 6 +-
.../Code/Numerics/FEM/itkFEMGenerateMesh.cxx | 14 +--
.../Numerics/FEM/itkFEMImageMetricLoad.txx | 16 +--
...MLoadImplementationGenericLandmarkLoad.cxx | 6 +-
.../FEM/itkFEMSolverCrankNicolson.cxx | 56 ++++-----
.../Numerics/FEM/itkFEMSolverCrankNicolson.h | 6 +-
.../Numerics/NeuralNetworks/CMakeLists.txt | 2 +-
.../NeuralNetworks/itkBackPropagationLayer.h | 8 +-
.../itkBackPropagationLayer.txx | 11 +-
.../itkBatchSupervisedTrainingFunction.h | 5 +-
.../itkBatchSupervisedTrainingFunction.txx | 21 +++-
.../itkCompletelyConnectedWeightSet.txx | 8 +-
.../NeuralNetworks/itkErrorFunctionBase.h | 11 +-
.../itkGaussianRadialBasisFunction.txx | 12 +-
.../itkIterativeSupervisedTrainingFunction.h | 5 +-
...itkIterativeSupervisedTrainingFunction.txx | 17 ++-
.../Numerics/NeuralNetworks/itkLayerBase.h | 22 +++-
.../Numerics/NeuralNetworks/itkLayerBase.txx | 24 +++-
.../itkLogSigmoidTransferFunction.txx | 6 +-
.../itkMeanSquaredErrorFunction.h | 7 +-
.../itkMeanSquaredErrorFunction.txx | 21 +++-
.../itkMultilayerNeuralNetworkBase.h | 21 ++--
.../itkMultilayerNeuralNetworkBase.txx | 36 +++++-
.../itkMultiquadricRadialBasisFunction.txx | 6 +-
.../NeuralNetworks/itkNeuralNetworkObject.h | 14 ++-
...eHiddenLayerBackPropagationNeuralNetwork.h | 10 +-
...iddenLayerBackPropagationNeuralNetwork.txx | 6 +-
.../itkRBFBackPropagationLearningFunction.txx | 6 +-
.../Numerics/NeuralNetworks/itkRBFLayer.h | 31 +++--
.../Numerics/NeuralNetworks/itkRBFLayer.txx | 78 ++++++++----
.../Numerics/NeuralNetworks/itkRBFNetwork.h | 17 ++-
.../Numerics/NeuralNetworks/itkRBFNetwork.txx | 34 +++++-
.../itkSigmoidTransferFunction.txx | 6 +-
.../itkSquaredDifferenceErrorFunction.h | 7 +-
.../itkSquaredDifferenceErrorFunction.txx | 20 +++-
.../itkSymmetricSigmoidTransferFunction.txx | 6 +-
.../itkTanHTransferFunction.txx | 10 +-
.../itkTanSigmoidTransferFunction.txx | 6 +-
.../NeuralNetworks/itkTrainingFunctionBase.h | 13 +-
.../itkTrainingFunctionBase.txx | 11 +-
...oHiddenLayerBackPropagationNeuralNetwork.h | 9 +-
...iddenLayerBackPropagationNeuralNetwork.txx | 6 +-
.../NeuralNetworks/itkWeightSetBase.h | 30 +++--
.../NeuralNetworks/itkWeightSetBase.txx | 25 ++--
.../Code/Numerics/Statistics/CMakeLists.txt | 7 +-
.../Statistics/itkEuclideanDistance.txx | 10 +-
...ationMaximizationMixtureModelEstimator.txx | 8 +-
.../Statistics/itkGaussianDensityFunction.txx | 8 +-
.../itkGaussianGoodnessOfFitComponent.txx | 8 +-
.../itkGaussianMixtureModelComponent.txx | 10 +-
.../itkGoodnessOfFitFunctionBase.txx | 6 +-
.../Code/Numerics/Statistics/itkHistogram.txx | 113 +++++-------------
.../itkHistogramToEntropyImageFilter.h | 8 +-
.../itkHistogramToLogProbabilityImageFilter.h | 8 +-
.../itkListSampleToHistogramGenerator.h | 11 +-
.../itkLogLikelihoodGoodnessOfFitFunction.txx | 6 +-
.../Statistics/itkNormalVariateGenerator.cxx | 18 +--
.../itkSampleToHistogramProjectionFilter.txx | 14 +--
.../itkScalarImageToHistogramGenerator.h | 12 +-
.../itkScalarImageToHistogramGenerator.txx | 28 ++++-
.../itkCumulativeGaussianCostFunction.cxx | 10 +-
.../itkCumulativeGaussianOptimizer.cxx | 16 +--
.../ITK/Code/Numerics/itkFRPROptimizer.cxx | 67 ++++++++---
.../ITK/Code/Numerics/itkFRPROptimizer.h | 22 +++-
...itkMultipleValuedNonLinearVnlOptimizer.cxx | 47 +++++++-
.../itkMultipleValuedNonLinearVnlOptimizer.h | 32 ++++-
...tkMultipleValuedVnlCostFunctionAdaptor.cxx | 69 ++++++++++-
.../itkMultipleValuedVnlCostFunctionAdaptor.h | 30 ++++-
.../itkOnePlusOneEvolutionaryOptimizer.cxx | 8 +-
.../ITK/Code/Numerics/itkSPSAOptimizer.cxx | 8 +-
.../ITK/Code/Numerics/itkSPSAOptimizer.h | 4 +
.../itkSingleValuedNonLinearVnlOptimizer.cxx | 7 +-
.../itkSingleValuedVnlCostFunctionAdaptor.cxx | 9 +-
78 files changed, 872 insertions(+), 460 deletions(-)
diff --git a/Utilities/ITK/Code/Numerics/CMakeLists.txt b/Utilities/ITK/Code/Numerics/CMakeLists.txt
index f0c12e6cf7..a56fbeb696 100644
--- a/Utilities/ITK/Code/Numerics/CMakeLists.txt
+++ b/Utilities/ITK/Code/Numerics/CMakeLists.txt
@@ -33,7 +33,10 @@ ADD_LIBRARY(ITKNumerics
lbfgsb.c
)
-TARGET_LINK_LIBRARIES (ITKNumerics ITKCommon)
+TARGET_LINK_LIBRARIES(ITKNumerics ITKCommon)
-INSTALL_TARGETS(/lib/otb ITKNumerics)
-INSTALL_FILES(/include/otb/Utilities/ITK/Numerics "(\\.h|\\.txx)$")
+IF(ITK_LIBRARY_PROPERTIES)
+ SET_TARGET_PROPERTIES(ITKNumerics PROPERTIES ${ITK_LIBRARY_PROPERTIES})
+ENDIF(ITK_LIBRARY_PROPERTIES)
+INSTALL_TARGETS(/lib/InsightToolkit ITKNumerics)
+INSTALL_FILES(/include/InsightToolkit/Numerics "(\\.h|\\.txx)$")
diff --git a/Utilities/ITK/Code/Numerics/FEM/CMakeLists.txt b/Utilities/ITK/Code/Numerics/FEM/CMakeLists.txt
index 20da66f0ea..e7da0a3797 100644
--- a/Utilities/ITK/Code/Numerics/FEM/CMakeLists.txt
+++ b/Utilities/ITK/Code/Numerics/FEM/CMakeLists.txt
@@ -188,5 +188,8 @@ IF(UNIX)
TARGET_LINK_LIBRARIES (ITKFEM -lm )
ENDIF(UNIX)
-INSTALL_TARGETS(/lib/otb ITKFEM)
-INSTALL_FILES(/include/otb/Utilities/ITK/Numerics/FEM "(\\.h|\\.txx)$")
+IF(ITK_LIBRARY_PROPERTIES)
+ SET_TARGET_PROPERTIES(ITKFEM PROPERTIES ${ITK_LIBRARY_PROPERTIES})
+ENDIF(ITK_LIBRARY_PROPERTIES)
+INSTALL_TARGETS(/lib/InsightToolkit ITKFEM)
+INSTALL_FILES(/include/InsightToolkit/Numerics/FEM "(\\.h|\\.txx)$")
diff --git a/Utilities/ITK/Code/Numerics/FEM/itkFEMElement2DC0LinearQuadrilateral.cxx b/Utilities/ITK/Code/Numerics/FEM/itkFEMElement2DC0LinearQuadrilateral.cxx
index 7b98e732d4..67fbfbbe57 100644
--- a/Utilities/ITK/Code/Numerics/FEM/itkFEMElement2DC0LinearQuadrilateral.cxx
+++ b/Utilities/ITK/Code/Numerics/FEM/itkFEMElement2DC0LinearQuadrilateral.cxx
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkFEMElement2DC0LinearQuadrilateral.cxx,v $
Language: C++
- Date: $Date: 2003/12/15 14:13:20 $
- Version: $Revision: 1.10 $
+ Date: $Date: 2006/03/19 04:37:20 $
+ Version: $Revision: 1.11 $
Copyright (c) Insight Software Consortium. All rights reserved.
See ITKCopyright.txt or http://www.itk.org/HTML/Copyright.htm for details.
@@ -173,7 +173,7 @@ Element2DC0LinearQuadrilateral
cn = (dx * ycn) - (dy * xcn);
localPt[0] = (2 * ce) / (-sqrt((be * be) - (2 * J1 * ce)) - be);
- localPt[1] = (2 * cn) / ( sqrt((bn * bn) + (2 * J2 * cn)) - bn);
+ localPt[1] = (2 * cn) / ( vcl_sqrt((bn * bn) + (2 * J2 * cn)) - bn);
bool isInside=true;
diff --git a/Utilities/ITK/Code/Numerics/FEM/itkFEMElement2DC1Beam.cxx b/Utilities/ITK/Code/Numerics/FEM/itkFEMElement2DC1Beam.cxx
index 615c4c8e66..7537699900 100644
--- a/Utilities/ITK/Code/Numerics/FEM/itkFEMElement2DC1Beam.cxx
+++ b/Utilities/ITK/Code/Numerics/FEM/itkFEMElement2DC1Beam.cxx
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkFEMElement2DC1Beam.cxx,v $
Language: C++
- Date: $Date: 2004/12/04 13:17:08 $
- Version: $Revision: 1.8 $
+ Date: $Date: 2006/03/19 04:37:20 $
+ Version: $Revision: 1.9 $
Copyright (c) Insight Software Consortium. All rights reserved.
See ITKCopyright.txt or http://www.itk.org/HTML/Copyright.htm for details.
@@ -148,7 +148,7 @@ MatrixType kb(NDOF,NDOF);
Float x=m_node[1]->GetCoordinates()[0]-m_node[0]->GetCoordinates()[0];
Float y=m_node[1]->GetCoordinates()[1]-m_node[0]->GetCoordinates()[1];
-Float l=sqrt(x*x+y*y);
+Float l=vcl_sqrt(x*x+y*y);
k[0][0]= 1; k[0][1]= 0; k[0][2]= 0; k[0][3]=-1; k[0][4]= 0; k[0][5]= 0;
k[1][0]= 0; k[1][1]= 0; k[1][2]= 0; k[1][3]= 0; k[1][4]= 0; k[1][5]= 0;
@@ -197,7 +197,7 @@ MatrixType k(NDOF,NDOF,0.0);
Float x=m_node[1]->GetCoordinates()[0]-m_node[0]->GetCoordinates()[0];
Float y=m_node[1]->GetCoordinates()[1]-m_node[0]->GetCoordinates()[1];
-Float l=sqrt(x*x+y*y);
+Float l=vcl_sqrt(x*x+y*y);
m[0][0]=2.0; m[0][3]=1.0;
m[3][0]=1.0; m[3][3]=2.0;
diff --git a/Utilities/ITK/Code/Numerics/FEM/itkFEMElement3DC0LinearHexahedron.cxx b/Utilities/ITK/Code/Numerics/FEM/itkFEMElement3DC0LinearHexahedron.cxx
index da7b228983..7395ea72cf 100644
--- a/Utilities/ITK/Code/Numerics/FEM/itkFEMElement3DC0LinearHexahedron.cxx
+++ b/Utilities/ITK/Code/Numerics/FEM/itkFEMElement3DC0LinearHexahedron.cxx
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkFEMElement3DC0LinearHexahedron.cxx,v $
Language: C++
- Date: $Date: 2003/12/15 14:13:20 $
- Version: $Revision: 1.6 $
+ Date: $Date: 2006/03/19 04:37:20 $
+ Version: $Revision: 1.7 $
Copyright (c) Insight Software Consortium. All rights reserved.
See ITKCopyright.txt or http://www.itk.org/HTML/Copyright.htm for details.
@@ -234,7 +234,7 @@ Element3DC0LinearHexahedron
// cn = (dx * ycn) - (dy * xcn);
// localPt[0] = (2 * ce) / (-sqrt((be * be) - (2 * J1 * ce)) - be);
-// localPt[1] = (2 * cn) / ( sqrt((bn * bn) + (2 * J2 * cn)) - bn);
+// localPt[1] = (2 * cn) / ( vcl_sqrt((bn * bn) + (2 * J2 * cn)) - bn);
// FIXME
bool IsInside=false;
diff --git a/Utilities/ITK/Code/Numerics/FEM/itkFEMElement3DC0LinearTetrahedron.cxx b/Utilities/ITK/Code/Numerics/FEM/itkFEMElement3DC0LinearTetrahedron.cxx
index 75b3a628f2..82c5f90151 100644
--- a/Utilities/ITK/Code/Numerics/FEM/itkFEMElement3DC0LinearTetrahedron.cxx
+++ b/Utilities/ITK/Code/Numerics/FEM/itkFEMElement3DC0LinearTetrahedron.cxx
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkFEMElement3DC0LinearTetrahedron.cxx,v $
Language: C++
- Date: $Date: 2004/07/22 15:52:10 $
- Version: $Revision: 1.7 $
+ Date: $Date: 2006/03/19 04:37:20 $
+ Version: $Revision: 1.8 $
Copyright (c) 2002 Insight Consortium. All rights reserved.
See ITKCopyright.txt or http://www.itk.org/HTML/Copyright.htm for details.
@@ -36,7 +36,7 @@ Element3DC0LinearTetrahedron
// FIXME: Write rules for other integration orders
pt.set_size(3);
- Float d = 1.0/sqrt(3.0);
+ Float d = 1.0/vcl_sqrt(3.0);
pt[0] = d;
pt[1] = d;
diff --git a/Utilities/ITK/Code/Numerics/FEM/itkFEMGenerateMesh.cxx b/Utilities/ITK/Code/Numerics/FEM/itkFEMGenerateMesh.cxx
index 77204a72f0..cb19b57a49 100644
--- a/Utilities/ITK/Code/Numerics/FEM/itkFEMGenerateMesh.cxx
+++ b/Utilities/ITK/Code/Numerics/FEM/itkFEMGenerateMesh.cxx
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkFEMGenerateMesh.cxx,v $
Language: C++
- Date: $Date: 2003/09/10 14:29:41 $
- Version: $Revision: 1.7 $
+ Date: $Date: 2006/03/19 04:37:20 $
+ Version: $Revision: 1.8 $
Copyright (c) Insight Software Consortium. All rights reserved.
See ITKCopyright.txt or http://www.itk.org/HTML/Copyright.htm for details.
@@ -44,8 +44,8 @@ void Generate2DRectilinearMesh(itk::fem::Element::ConstPointer e0, Solver& S, vn
S.node.clear();
// Number of nodes in each dimension
- Nel[0]=floor(Nel[0]);
- Nel[1]=floor(Nel[1]);
+ Nel[0]=vcl_floor(Nel[0]);
+ Nel[1]=vcl_floor(Nel[1]);
double Ni=static_cast<double>(Nel[0]);
double Nj=static_cast<double>(Nel[1]);
@@ -101,9 +101,9 @@ void Generate3DRectilinearMesh
}
// Number of nodes in each dimension
- Nel[0]=floor(Nel[0]);
- Nel[1]=floor(Nel[1]);
- Nel[2]=floor(Nel[2]);
+ Nel[0]=vcl_floor(Nel[0]);
+ Nel[1]=vcl_floor(Nel[1]);
+ Nel[2]=vcl_floor(Nel[2]);
double Ni=static_cast<double>(Nel[0]);
double Nj=static_cast<double>(Nel[1]);
double Nk=static_cast<double>(Nel[2]);
diff --git a/Utilities/ITK/Code/Numerics/FEM/itkFEMImageMetricLoad.txx b/Utilities/ITK/Code/Numerics/FEM/itkFEMImageMetricLoad.txx
index fe228c496c..fedd1908fe 100644
--- a/Utilities/ITK/Code/Numerics/FEM/itkFEMImageMetricLoad.txx
+++ b/Utilities/ITK/Code/Numerics/FEM/itkFEMImageMetricLoad.txx
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit (ITK)
Module: $RCSfile: itkFEMImageMetricLoad.txx,v $
Language: C++
- Date: $Date: 2003/12/15 14:13:20 $
- Version: $Revision: 1.30 $
+ Date: $Date: 2006/03/19 04:37:20 $
+ Version: $Revision: 1.31 $
=========================================================================*/
#ifndef _itkFEMImageMetricLoad_txx_
@@ -146,7 +146,7 @@ ImageMetricLoad<TMoving , TFixed>::EvaluateMetricGivenSolution( Element::ArrayTy
float tempe=0.0;
try
{
- tempe=fabs(GetMetric(InVec));
+ tempe=vcl_fabs(GetMetric(InVec));
}
catch( itk::ExceptionObject & )
{
@@ -164,7 +164,7 @@ ImageMetricLoad<TMoving , TFixed>::EvaluateMetricGivenSolution( Element::ArrayTy
}
//std::cout << " def e " << defe << " sim e " << energy*m_Gamma << std::endl;
- return fabs((double)energy*(double)m_Gamma-(double)defe);
+ return vcl_fabs((double)energy*(double)m_Gamma-(double)defe);
}
@@ -192,7 +192,7 @@ ImageMetricLoad<TMoving , TFixed>::Fe
for( unsigned int k = 0; k < ImageDimension; k++ ) {
if ( vnl_math_isnan(Gpos[k]) || vnl_math_isinf(Gpos[k]) ||
vnl_math_isnan(Gsol[k]) || vnl_math_isinf(Gsol[k]) ||
- fabs(Gpos[k]) > 1.e33 || fabs(Gsol[k]) > 1.e33 )
+ vcl_fabs(Gpos[k]) > 1.e33 || vcl_fabs(Gsol[k]) > 1.e33 )
{
OutVec.set_size(ImageDimension); OutVec.fill(0.0); return OutVec;
}
@@ -201,7 +201,7 @@ ImageMetricLoad<TMoving , TFixed>::Fe
// OutVec=this->GetPolynomialFitToMetric(Gpos,Gsol); // gradient direction
// for( unsigned int k = 0; k < ImageDimension; k++ ) {
// if ( vnl_math_isnan(OutVec[k]) || vnl_math_isinf(OutVec[k])
-// || fabs(OutVec[k]) > 1.e33 ) OutVec[k]=0.0;
+// || vcl_fabs(OutVec[k]) > 1.e33 ) OutVec[k]=0.0;
// else OutVec[k]=m_Sign*OutVec[k];
// }
// return OutVec;
@@ -270,9 +270,9 @@ ImageMetricLoad<TMoving , TFixed>::Fe
// IN FACT, IT SEEMS MEANSQRS AND NCC POINT IN DIFFT DIRS
//std::cout << " deriv " << derivative << " val " << measure << endl;
//if (m_Temp !=0.0)
- //return OutVec * exp(-1.*OutVec.magnitude()/m_Temp);
+ //return OutVec * vcl_exp(-1.*OutVec.magnitude()/m_Temp);
//else
- return OutVec/sqrt(gmag);
+ return OutVec/vcl_sqrt(gmag);
}
diff --git a/Utilities/ITK/Code/Numerics/FEM/itkFEMLoadImplementationGenericLandmarkLoad.cxx b/Utilities/ITK/Code/Numerics/FEM/itkFEMLoadImplementationGenericLandmarkLoad.cxx
index 18ff538e33..ee2ab356c8 100644
--- a/Utilities/ITK/Code/Numerics/FEM/itkFEMLoadImplementationGenericLandmarkLoad.cxx
+++ b/Utilities/ITK/Code/Numerics/FEM/itkFEMLoadImplementationGenericLandmarkLoad.cxx
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkFEMLoadImplementationGenericLandmarkLoad.cxx,v $
Language: C++
- Date: $Date: 2003/12/15 14:13:20 $
- Version: $Revision: 1.8 $
+ Date: $Date: 2006/03/19 04:37:20 $
+ Version: $Revision: 1.9 $
Copyright (c) Insight Software Consortium. All rights reserved.
See ITKCopyright.txt or http://www.itk.org/HTML/Copyright.htm for details.
@@ -64,7 +64,7 @@ LoadImplementationGenericLandmarkLoad
force = (load->m_target-new_source) / load->eta ;
// std::cout << " disp " << disp << std::endl;
- //force /= sqrt(fmag);
+ //force /= vcl_sqrt(fmag);
new_source = (load->GetTarget() - new_source);
// std::cout << " force = " << force << " distance " << new_source.magnitude() << std::endl;
diff --git a/Utilities/ITK/Code/Numerics/FEM/itkFEMSolverCrankNicolson.cxx b/Utilities/ITK/Code/Numerics/FEM/itkFEMSolverCrankNicolson.cxx
index 14f1197a8f..76510b89d2 100644
--- a/Utilities/ITK/Code/Numerics/FEM/itkFEMSolverCrankNicolson.cxx
+++ b/Utilities/ITK/Code/Numerics/FEM/itkFEMSolverCrankNicolson.cxx
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkFEMSolverCrankNicolson.cxx,v $
Language: C++
- Date: $Date: 2005/11/17 22:07:13 $
- Version: $Revision: 1.37 $
+ Date: $Date: 2006/03/19 04:37:20 $
+ Version: $Revision: 1.38 $
Copyright (c) Insight Software Consortium. All rights reserved.
See ITKCopyright.txt or http://www.itk.org/HTML/Copyright.htm for details.
@@ -269,8 +269,8 @@ void SolverCrankNicolson::FindBracketingTriplet(Float* a, Float* b, Float* c)
Float ax, bx,cx;
ax=0.0; bx=1.;
Float fc;
- Float fa=fabs(EvaluateResidual(ax));
- Float fb=fabs(EvaluateResidual(bx));
+ Float fa=vcl_fabs(EvaluateResidual(ax));
+ Float fb=vcl_fabs(EvaluateResidual(bx));
Float ulim,u,r,q,fu,dum;
@@ -281,10 +281,10 @@ void SolverCrankNicolson::FindBracketingTriplet(Float* a, Float* b, Float* c)
}
cx=bx+Gold*(bx-ax); // first guess for c - the 3rd pt needed to bracket the min
- fc=fabs(EvaluateResidual(cx));
+ fc=vcl_fabs(EvaluateResidual(cx));
- while (fb > fc /*&& fabs(ax) < 3. && fabs(bx) < 3. && fabs(cx) < 3.*/)
+ while (fb > fc /*&& vcl_fabs(ax) < 3. && vcl_fabs(bx) < 3. && vcl_fabs(cx) < 3.*/)
{
r=(bx-ax)*(fb-fc);
q=(bx-cx)*(fb-fa);
@@ -293,7 +293,7 @@ void SolverCrankNicolson::FindBracketingTriplet(Float* a, Float* b, Float* c)
ulim=bx + Glimit*(cx-bx);
if ((bx-u)*(u-cx) > 0.0)
{
- fu=fabs(EvaluateResidual(u));
+ fu=vcl_fabs(EvaluateResidual(u));
if (fu < fc)
{
ax=bx;
@@ -309,28 +309,28 @@ void SolverCrankNicolson::FindBracketingTriplet(Float* a, Float* b, Float* c)
}
u=cx+Gold*(cx-bx);
- fu=fabs(EvaluateResidual(u));
+ fu=vcl_fabs(EvaluateResidual(u));
}
else if ( (cx-u)*(u-ulim) > 0.0)
{
- fu=fabs(EvaluateResidual(u));
+ fu=vcl_fabs(EvaluateResidual(u));
if (fu < fc)
{
bx=cx; cx=u; u=cx+Gold*(cx-bx);
- fb=fc; fc=fu; fu=fabs(EvaluateResidual(u));
+ fb=fc; fc=fu; fu=vcl_fabs(EvaluateResidual(u));
}
}
else if ( (u-ulim)*(ulim-cx) >= 0.0)
{
u=ulim;
- fu=fabs(EvaluateResidual(u));
+ fu=vcl_fabs(EvaluateResidual(u));
}
else
{
u=cx+Gold*(cx-bx);
- fu=fabs(EvaluateResidual(u));
+ fu=vcl_fabs(EvaluateResidual(u));
}
ax=bx; bx=cx; cx=u;
@@ -338,7 +338,7 @@ void SolverCrankNicolson::FindBracketingTriplet(Float* a, Float* b, Float* c)
}
- if ( fabs(ax) > 1.e3 || fabs(bx) > 1.e3 || fabs(cx) > 1.e3)
+ if ( vcl_fabs(ax) > 1.e3 || vcl_fabs(bx) > 1.e3 || vcl_fabs(cx) > 1.e3)
{ ax=-2.0; bx=1.0; cx=2.0; } // to avoid crazy numbers caused by bad bracket (u goes nuts)
*a=ax; *b=bx; *c=cx;
@@ -368,12 +368,12 @@ Element::Float SolverCrankNicolson::BrentsMethod(Float tol,unsigned int MaxIters
b=((ax > cx) ? ax : cx);
x=w=v=bx;
- fw=fv=fx=fabs(EvaluateResidual(x));
+ fw=fv=fx=vcl_fabs(EvaluateResidual(x));
for (iter = 1; iter <=MaxIters; iter++)
{
xm=0.5*(a+b);
- tol2=2.0*(tol1=tol*fabs(x)+ZEPS);
+ tol2=2.0*(tol1=tol*vcl_fabs(x)+ZEPS);
if (fabs(x-xm) <= (tol2-0.5*(b-a)))
{
xmin=x;
@@ -387,10 +387,10 @@ Element::Float SolverCrankNicolson::BrentsMethod(Float tol,unsigned int MaxIters
p=(x-v)*q-(x-w)*r;
q=2.0*(q-r);
if (q>0.0) p = -1.*p;
- q=fabs(q);
+ q=vcl_fabs(q);
etemp=e;
e=d;
- if (fabs(p) >= fabs(0.5*q*etemp) || p <= q*(a-x) || p >= q*(b-x))
+ if (fabs(p) >= vcl_fabs(0.5*q*etemp) || p <= q*(a-x) || p >= q*(b-x))
d=CGOLD*(e=(x>=xm ? a-x : b-x));
else{
if (q == 0.0) q=q +ZEPS;
@@ -403,7 +403,7 @@ Element::Float SolverCrankNicolson::BrentsMethod(Float tol,unsigned int MaxIters
}
u=(fabs(d) >= tol1 ? x+d : x + GSSign(tol1,d));
- fu=fabs(EvaluateResidual(u));
+ fu=vcl_fabs(EvaluateResidual(u));
if (fu <= fx){
if ( u >= x ) a=x; else b=x;
v=w; w=x;x=u;
@@ -446,25 +446,25 @@ Element::Float SolverCrankNicolson::GoldenSection(Float tol,unsigned int MaxIter
x0=ax;
x3=cx;
- if (fabs(cx-bx) > fabs(bx-ax)){
+ if (fabs(cx-bx) > vcl_fabs(bx-ax)){
x1=bx;
x2=bx+C*(cx-bx);
} else {
x2=bx;
x1=bx-C*(bx-ax);
}
- f1=fabs(EvaluateResidual(x1));
- f2=fabs(EvaluateResidual(x2));
+ f1=vcl_fabs(EvaluateResidual(x1));
+ f2=vcl_fabs(EvaluateResidual(x2));
unsigned int iters=0;
- while (fabs(x3-x0) > tol*(fabs(x1)+fabs(x2)) && iters < MaxIters)
+ while (fabs(x3-x0) > tol*(fabs(x1)+vcl_fabs(x2)) && iters < MaxIters)
{
iters++;
if (f2 < f1){
x0=x1; x1=x2; x2=R*x1+C*x3;
- f1=f2; f2=fabs(EvaluateResidual(x2));
+ f1=f2; f2=vcl_fabs(EvaluateResidual(x2));
} else {
x3=x2; x2=x1; x1=R*x2+C*x0;
- f2=f1; f1=fabs(EvaluateResidual(x1));
+ f2=f1; f1=vcl_fabs(EvaluateResidual(x1));
}
}
if (f1<f2){
@@ -581,7 +581,7 @@ Element::Float SolverCrankNicolson::EvaluateResidual(Float t)
}
DeformationEnergy+=iSolVal*TempRowVal;
}
- Float Energy=(Float) fabs(DeformationEnergy-ForceEnergy);
+ Float Energy=(Float) vcl_fabs(DeformationEnergy-ForceEnergy);
return Energy;
}
@@ -610,7 +610,7 @@ void SolverCrankNicolson::AddToDisplacements(Float optimum)
else if (CurrentSolution > maxs2 ) {
maxs2=CurrentSolution;
}
- if (fabs(CurrentSolution) > absmax) absmax=fabs(CurrentSolution);
+ if (fabs(CurrentSolution) > absmax) absmax=vcl_fabs(CurrentSolution);
// note: set rather than add - i.e. last solution of system not total solution
#ifdef LOCE
@@ -634,8 +634,8 @@ void SolverCrankNicolson::AddToDisplacements(Float optimum)
m_ls->AddVectorValue(i , CurrentForce, ForceTotalIndex);
CurrentTotSolution=m_ls->GetSolutionValue(i,TotalSolutionIndex);
- if ( fabs(CurrentTotSolution) > maxs ) {
- maxs=fabs(CurrentTotSolution);
+ if ( vcl_fabs(CurrentTotSolution) > maxs ) {
+ maxs=vcl_fabs(CurrentTotSolution);
}
diff --git a/Utilities/ITK/Code/Numerics/FEM/itkFEMSolverCrankNicolson.h b/Utilities/ITK/Code/Numerics/FEM/itkFEMSolverCrankNicolson.h
index e06e1fc7c5..ec3aee735d 100644
--- a/Utilities/ITK/Code/Numerics/FEM/itkFEMSolverCrankNicolson.h
+++ b/Utilities/ITK/Code/Numerics/FEM/itkFEMSolverCrankNicolson.h
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkFEMSolverCrankNicolson.h,v $
Language: C++
- Date: $Date: 2003/09/10 14:29:44 $
- Version: $Revision: 1.17 $
+ Date: $Date: 2006/03/19 04:37:20 $
+ Version: $Revision: 1.18 $
Copyright (c) Insight Software Consortium. All rights reserved.
See ITKCopyright.txt or http://www.itk.org/HTML/Copyright.htm for details.
@@ -122,7 +122,7 @@ public:
Float BrentsMethod(Float tol=0.01,unsigned int MaxIters=25);
Float EvaluateResidual(Float t=1.0);
Float GetDeformationEnergy(Float t=1.0);
- inline Float GSSign(Float a,Float b) { return (b > 0.0 ? fabs(a) : -1.*fabs(a)); }
+ inline Float GSSign(Float a,Float b) { return (b > 0.0 ? vcl_fabs(a) : -1.*vcl_fabs(a)); }
inline Float GSMax(Float a,Float b) { return (a > b ? a : b); }
void SetEnergyToMin(Float xmin);
diff --git a/Utilities/ITK/Code/Numerics/NeuralNetworks/CMakeLists.txt b/Utilities/ITK/Code/Numerics/NeuralNetworks/CMakeLists.txt
index 00b40cc77d..3230606db2 100644
--- a/Utilities/ITK/Code/Numerics/NeuralNetworks/CMakeLists.txt
+++ b/Utilities/ITK/Code/Numerics/NeuralNetworks/CMakeLists.txt
@@ -1 +1 @@
-INSTALL_FILES(/include/otb/Utilities/ITK/Numerics/NeuralNetworks "(\\.h|\\.txx)$")
+INSTALL_FILES(/include/InsightToolkit/Numerics/NeuralNetworks "(\\.h|\\.txx)$")
diff --git a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkBackPropagationLayer.h b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkBackPropagationLayer.h
index 9d6c825f84..327ab56037 100644
--- a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkBackPropagationLayer.h
+++ b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkBackPropagationLayer.h
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkBackPropagationLayer.h,v $
Language: C++
- Date: $Date: 2005/08/02 19:17:37 $
- Version: $Revision: 1.1 $
+ Date: $Date: 2006/04/17 21:34:31 $
+ Version: $Revision: 1.3 $
Copyright (c) Insight Software Consortium. All rights reserved.
See ITKCopyright.txt or http://www.itk.org/HTML/Copyright.htm for details.
@@ -41,7 +41,7 @@ public:
typedef typename Superclass::ValueType ValueType;
typedef typename Superclass::ValuePointer ValuePointer;
typedef vnl_vector<ValueType> NodeVectorType;
-
+ typedef typename Superclass::InternalVectorType InternalVectorType;
typedef typename Superclass::OutputVectorType OutputVectorType;
//Member Functions
@@ -58,7 +58,7 @@ public:
void ForwardPropagate();
void ForwardPropagate(TVector);
- void BackwardPropagate(TOutput e);
+ void BackwardPropagate(InternalVectorType e);
void BackwardPropagate();
void SetOutputErrorValues(TOutput);
diff --git a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkBackPropagationLayer.txx b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkBackPropagationLayer.txx
index ceeab19507..cf9511d020 100644
--- a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkBackPropagationLayer.txx
+++ b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkBackPropagationLayer.txx
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkBackPropagationLayer.txx,v $
Language: C++
- Date: $Date: 2005/08/09 20:27:15 $
- Version: $Revision: 1.4 $
+ Date: $Date: 2006/04/17 19:34:45 $
+ 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.
@@ -210,7 +210,7 @@ BackPropagationLayer<TVector,TOutput>
template<class TVector, class TOutput>
void
BackPropagationLayer<TVector,TOutput>
-::BackwardPropagate(TOutput errors)
+::BackwardPropagate(InternalVectorType errors)
{
int num_nodes = this->GetNumberOfNodes();
typename Superclass::WeightSetType::Pointer inputweightset;
@@ -253,7 +253,10 @@ void
BackPropagationLayer<TVector,TOutput>
::SetOutputErrorValues(TOutput errors)
{
- m_OutputErrorValues = errors.GetVnlVector();
+ for(unsigned int i=0; i<errors.Size(); i++)
+ m_OutputErrorValues[i] = errors[i];
+
+ //m_OutputErrorValues = errors.GetVnlVector();
this->Modified();
}
diff --git a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkBatchSupervisedTrainingFunction.h b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkBatchSupervisedTrainingFunction.h
index a80e3a498e..d0541a17de 100644
--- a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkBatchSupervisedTrainingFunction.h
+++ b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkBatchSupervisedTrainingFunction.h
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkBatchSupervisedTrainingFunction.h,v $
Language: C++
- Date: $Date: 2005/08/02 19:17:37 $
- Version: $Revision: 1.1 $
+ Date: $Date: 2006/04/18 11:23:29 $
+ Version: $Revision: 1.2 $
Copyright (c) Insight Software Consortium. All rights reserved.
See ITKCopyright.txt or http://www.itk.org/HTML/Copyright.htm for details.
@@ -41,6 +41,7 @@ public:
itkNewMacro(Self);
typedef typename Superclass::NetworkType NetworkType;
+ typedef typename Superclass::InternalVectorType InternalVectorType;
/** Set the number of iterations */
void SetNumOfIterations(long i);
diff --git a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkBatchSupervisedTrainingFunction.txx b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkBatchSupervisedTrainingFunction.txx
index 7e21621493..2bbf27a2cd 100644
--- a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkBatchSupervisedTrainingFunction.txx
+++ b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkBatchSupervisedTrainingFunction.txx
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkBatchSupervisedTrainingFunction.txx,v $
Language: C++
- Date: $Date: 2005/08/03 12:06:12 $
- Version: $Revision: 1.2 $
+ Date: $Date: 2006/04/17 19:34:45 $
+ Version: $Revision: 1.3 $
Copyright (c) Insight Software Consortium. All rights reserved.
See ITKCopyright.txt or http://www.itk.org/HTML/Copyright.htm for details.
@@ -50,10 +50,16 @@ void BatchSupervisedTrainingFunction<TSample,TOutput,ScalarType>
{
this->SetTrainingSamples(samples);
this->SetTargetValues(targets);
- typename Superclass::OutputVectorType outputvector;
+
+ InternalVectorType outputvector;
+ InternalVectorType errorvector;
+ outputvector.SetSize(targets->GetMeasurementVectorSize());
+ errorvector.SetSize(targets->GetMeasurementVectorSize());
+ std::cout<<"Target dim ="<<targets->GetMeasurementVectorSize()<<std::endl;
+ //typename Superclass::OutputVectorType outputvector;
typename Superclass::VectorType inputvector;
typename Superclass::OutputVectorType targetvector;
- typename Superclass::OutputVectorType errorvector;
+ //typename Superclass::OutputVectorType errorvector;
long num_iterations = this->GetIterations();
m_Stop = false;
@@ -67,8 +73,11 @@ void BatchSupervisedTrainingFunction<TSample,TOutput,ScalarType>
{
inputvector = this->m_InputSamples[i];
targetvector = this->m_Targets[i];
- outputvector = net->GenerateOutput(inputvector);
- errorvector = targetvector - outputvector;
+
+ outputvector=net->GenerateOutput(inputvector);
+
+ for(unsigned int k=0; k<targetvector.Size(); k++)
+ errorvector[k] = targetvector[k] - outputvector[k];
outf << this->m_PerformanceFunction->Evaluate(errorvector) <<
" "<<errorvector[0]<<std::endl;
diff --git a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkCompletelyConnectedWeightSet.txx b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkCompletelyConnectedWeightSet.txx
index 433342db65..14df086197 100644
--- a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkCompletelyConnectedWeightSet.txx
+++ b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkCompletelyConnectedWeightSet.txx
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkCompletelyConnectedWeightSet.txx,v $
Language: C++
- Date: $Date: 2005/08/02 19:17:37 $
- Version: $Revision: 1.1 $
+ Date: $Date: 2006/04/17 19:35:36 $
+ Version: $Revision: 1.2 $
Copyright (c) Insight Software Consortium. All rights reserved.
See ITKCopyright.txt or http://www.itk.org/HTML/Copyright.htm for details.
@@ -37,8 +37,8 @@ CompletelyConnectedWeightSet<TVector,TOutput>
::SetCompleteConnectivity()
{
vnl_matrix<int> c;
- unsigned int rows = WeightSetBase<TVector, TOutput>::GetNumberOfInputNodes();
- unsigned int cols = WeightSetBase<TVector, TOutput>::GetNumberOfOutputNodes();
+ unsigned int rows = WeightSetBase<TVector, TOutput>::GetNumberOfOutputNodes();
+ unsigned int cols = WeightSetBase<TVector, TOutput>::GetNumberOfInputNodes();
std::cout << "Connectivity matrix size= " << rows << " " << cols << std::endl;
c.set_size(rows, cols);
c.fill(1);
diff --git a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkErrorFunctionBase.h b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkErrorFunctionBase.h
index 0f3c4c31e6..2580f2d76c 100644
--- a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkErrorFunctionBase.h
+++ b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkErrorFunctionBase.h
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkErrorFunctionBase.h,v $
Language: C++
- Date: $Date: 2005/08/02 19:17:37 $
- Version: $Revision: 1.1 $
+ Date: $Date: 2006/04/17 19:34:45 $
+ Version: $Revision: 1.2 $
Copyright (c) Insight Software Consortium. All rights reserved.
See ITKCopyright.txt or http://www.itk.org/HTML/Copyright.htm for details.
@@ -44,10 +44,13 @@ public:
/** Output type */
typedef TOutput OutputType;
- virtual OutputType Evaluate(const ErrorVectorType&) const = 0;
+ typedef Array<double> InternalVectorType;
- virtual ErrorVectorType EvaluateDerivative(const ErrorVectorType&) const = 0;
+ virtual OutputType Evaluate(const ErrorVectorType&) const = 0;
+ //virtual ErrorVectorType EvaluateDerivative(const ErrorVectorType&) const = 0;
+ virtual InternalVectorType EvaluateDerivative(const ErrorVectorType&) const = 0;
+
protected:
ErrorFunctionBase(){};
diff --git a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkGaussianRadialBasisFunction.txx b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkGaussianRadialBasisFunction.txx
index 74c6177426..db6185b507 100644
--- a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkGaussianRadialBasisFunction.txx
+++ b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkGaussianRadialBasisFunction.txx
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkGaussianRadialBasisFunction.txx,v $
Language: C++
- Date: $Date: 2005/08/03 15:47:06 $
- Version: $Revision: 1.2 $
+ Date: $Date: 2006/04/18 11:23:29 $
+ Version: $Revision: 1.4 $
Copyright (c) Insight Software Consortium. All rights reserved.
See ITKCopyright.txt or http://www.itk.org/HTML/Copyright.htm for details.
@@ -49,7 +49,7 @@ GaussianRadialBasisFunction< ScalarType>
{
ScalarType val;
ScalarType radius = Superclass::GetRadius();
- val = exp(-0.5*pow(input,2)/pow(radius,2));
+ val = vcl_exp(-0.5*vcl_pow(input,2)/vcl_pow(radius,2));
return val;
}
@@ -60,18 +60,18 @@ GaussianRadialBasisFunction< ScalarType>
::EvaluateDerivative(const ScalarType& dist,const ArrayType& input,
char mode,int element_id) const
{
- ScalarType val;
+ ScalarType val = 0;;
ScalarType radius=Superclass::GetRadius();
ArrayType center = Superclass::GetCenter();
if(mode=='u') //w.r.t centers
{
- ScalarType temp1= pow(radius,2);
+ ScalarType temp1= vcl_pow(radius,2);
val=Evaluate(dist)
*(input.GetElement(element_id)-center.GetElement(element_id))/temp1;
}
else if(mode=='s') // w.r.t radius
{
- val=Evaluate(dist)*pow(dist,2)/pow(radius,3);
+ val=Evaluate(dist)*vcl_pow(dist,2)/vcl_pow(radius,3);
}
return val;
}
diff --git a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkIterativeSupervisedTrainingFunction.h b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkIterativeSupervisedTrainingFunction.h
index 52f3c368d1..200067c988 100644
--- a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkIterativeSupervisedTrainingFunction.h
+++ b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkIterativeSupervisedTrainingFunction.h
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkIterativeSupervisedTrainingFunction.h,v $
Language: C++
- Date: $Date: 2005/08/02 19:17:37 $
- Version: $Revision: 1.1 $
+ Date: $Date: 2006/04/18 11:23:29 $
+ Version: $Revision: 1.2 $
Copyright (c) Insight Software Consortium. All rights reserved.
See ITKCopyright.txt or http://www.itk.org/HTML/Copyright.htm for details.
@@ -43,6 +43,7 @@ public:
itkNewMacro(Self);
typedef typename Superclass::NetworkType NetworkType;
+ typedef typename Superclass::InternalVectorType InternalVectorType;
void SetNumOfIterations(long i);
diff --git a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkIterativeSupervisedTrainingFunction.txx b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkIterativeSupervisedTrainingFunction.txx
index c151e92884..2d98aadfde 100644
--- a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkIterativeSupervisedTrainingFunction.txx
+++ b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkIterativeSupervisedTrainingFunction.txx
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkIterativeSupervisedTrainingFunction.txx,v $
Language: C++
- Date: $Date: 2005/08/02 19:17:37 $
- Version: $Revision: 1.1 $
+ Date: $Date: 2006/04/17 19:34:45 $
+ Version: $Revision: 1.2 $
Copyright (c) Insight Software Consortium. All rights reserved.
See ITKCopyright.txt or http://www.itk.org/HTML/Copyright.htm for details.
@@ -50,10 +50,16 @@ void IterativeSupervisedTrainingFunction<TSample,TOutput,ScalarType>
{
this->SetTrainingSamples(samples);
this->SetTargetValues(targets);
- typename Superclass::OutputVectorType outputvector;
+
+ InternalVectorType outputvector;
+ InternalVectorType errorvector;
+ outputvector.SetSize(targets->GetMeasurementVectorSize());
+ errorvector.SetSize(targets->GetMeasurementVectorSize());
+
+ //typename Superclass::OutputVectorType outputvector;
typename Superclass::VectorType inputvector;
typename Superclass::OutputVectorType targetvector;
- typename Superclass::OutputVectorType errorvector;
+ //typename Superclass::OutputVectorType errorvector;
std::ofstream outfile;
outfile.open("output.txt");
@@ -67,7 +73,8 @@ void IterativeSupervisedTrainingFunction<TSample,TOutput,ScalarType>
inputvector = this->m_InputSamples[temp];
targetvector = this->m_Targets[temp];
outputvector = Net->GenerateOutput(inputvector);
- errorvector = targetvector - outputvector;
+ for(unsigned int k=0; k<targetvector.Size(); k++)
+ errorvector[k] = targetvector[k] - outputvector[k];
outfile <<errorvector[0] << std::endl;
Net->BackwardPropagate(this->m_PerformanceFunction->EvaluateDerivative(errorvector));
Net->UpdateWeights(this->m_LearningRate);
diff --git a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkLayerBase.h b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkLayerBase.h
index 887ef730c6..5844947ce8 100644
--- a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkLayerBase.h
+++ b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkLayerBase.h
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkLayerBase.h,v $
Language: C++
- Date: $Date: 2005/08/02 19:17:37 $
- Version: $Revision: 1.1 $
+ Date: $Date: 2006/04/17 19:34:45 $
+ Version: $Revision: 1.2 $
Copyright (c) Insight Software Consortium. All rights reserved.
See ITKCopyright.txt or http://www.itk.org/HTML/Copyright.htm for details.
@@ -51,6 +51,7 @@ public:
typedef typename TVector::ValueType ValueType;
typedef ValueType* ValuePointer;
typedef vnl_vector<ValueType> NodeVectorType;
+ typedef Array<ValueType> InternalVectorType;
typedef WeightSetBase<TVector,TOutput> WeightSetType;
@@ -62,6 +63,8 @@ public:
typedef typename TransferFunctionType::Pointer TransferFunctionPointer;
+ typedef typename WeightSetType::Pointer WeightSetPointer;
+
virtual void SetNumberOfNodes(unsigned int);
unsigned int GetNumberOfNodes();
@@ -73,7 +76,7 @@ public:
virtual void ForwardPropagate(TVector){};
- virtual void BackwardPropagate(TOutput){};
+ virtual void BackwardPropagate(InternalVectorType){};
virtual void BackwardPropagate(){};
virtual ValueType GetOutputErrorValue(unsigned int) = 0;
@@ -83,10 +86,12 @@ public:
virtual ValuePointer GetInputErrorVector() = 0;
virtual void SetInputErrorValue(ValueType, int) {};
- itkSetObjectMacro(InputWeightSet, WeightSetType);
+ //itkSetObjectMacro(InputWeightSet, WeightSetType);
+ void SetInputWeightSet(WeightSetType*);
itkGetObjectMacro(InputWeightSet, WeightSetType);
- itkSetObjectMacro(OutputWeightSet, WeightSetType);
+ //itkSetObjectMacro(OutputWeightSet, WeightSetType);
+ void SetOutputWeightSet(WeightSetType*);
itkGetObjectMacro(OutputWeightSet, WeightSetType);
void SetNodeInputFunction(InputFunctionType* f);
@@ -101,9 +106,13 @@ public:
itkSetMacro(LayerType, unsigned int);
itkGetMacro(LayerType, unsigned int);
+ itkSetMacro(LayerId,int);
+ itkGetMacro(LayerId,int);
+
virtual void SetBias(ValueType) = 0;
virtual ValueType GetBias() = 0;
+
protected:
LayerBase();
@@ -113,6 +122,7 @@ protected:
virtual void PrintSelf( std::ostream& os, Indent indent ) const;
unsigned int m_LayerType; //input, hidden, output
+ unsigned int m_LayerId; //input, hidden, output
unsigned int m_NumberOfNodes;
typename WeightSetType::Pointer m_InputWeightSet;
@@ -120,6 +130,8 @@ protected:
TransferFunctionPointer m_ActivationFunction;
InputFunctionPointer m_NodeInputFunction;
+
+
}; //class layer base
diff --git a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkLayerBase.txx b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkLayerBase.txx
index c5cfc34e75..b26796301d 100644
--- a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkLayerBase.txx
+++ b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkLayerBase.txx
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkLayerBase.txx,v $
Language: C++
- Date: $Date: 2005/08/02 19:17:37 $
- Version: $Revision: 1.1 $
+ Date: $Date: 2006/04/17 19:34:45 $
+ Version: $Revision: 1.2 $
Copyright (c) Insight Software Consortium. All rights reserved.
See ITKCopyright.txt or http://www.itk.org/HTML/Copyright.htm for details.
@@ -72,6 +72,26 @@ LayerBase<TVector,TOutput>
this->Modified();
}
+template<class TVector, class TOutput>
+void
+LayerBase<TVector,TOutput>
+::SetInputWeightSet(WeightSetType* weightset)
+{
+ m_InputWeightSet=weightset;
+ m_InputWeightSet->SetOutputLayerId(m_LayerId);
+ this->Modified();
+}
+
+template<class TVector, class TOutput>
+void
+LayerBase<TVector,TOutput>
+::SetOutputWeightSet(WeightSetType* weightset)
+{
+ m_OutputWeightSet=weightset;
+ m_OutputWeightSet->SetInputLayerId(m_LayerId);
+ this->Modified();
+}
+
/** Print the object */
template<class TVector, class TOutput>
void
diff --git a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkLogSigmoidTransferFunction.txx b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkLogSigmoidTransferFunction.txx
index f88eb1b570..91faf66b94 100644
--- a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkLogSigmoidTransferFunction.txx
+++ b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkLogSigmoidTransferFunction.txx
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkLogSigmoidTransferFunction.txx,v $
Language: C++
- Date: $Date: 2005/08/02 19:17:37 $
- Version: $Revision: 1.1 $
+ Date: $Date: 2006/03/19 04:37:20 $
+ Version: $Revision: 1.2 $
Copyright (c) Insight Software Consortium. All rights reserved.
See ITKCopyright.txt or http://www.itk.org/HTML/Copyright.htm for details.
@@ -45,7 +45,7 @@ ScalarType
LogSigmoidTransferFunction<ScalarType>
::Evaluate(const ScalarType& input) const
{
- const ScalarType v = 1.0 / (1.0 + exp(-input));
+ const ScalarType v = 1.0 / (1.0 + vcl_exp(-input));
return v;
}
diff --git a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkMeanSquaredErrorFunction.h b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkMeanSquaredErrorFunction.h
index 9d235ce5c1..244f70204e 100644
--- a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkMeanSquaredErrorFunction.h
+++ b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkMeanSquaredErrorFunction.h
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkMeanSquaredErrorFunction.h,v $
Language: C++
- Date: $Date: 2005/08/02 19:17:37 $
- Version: $Revision: 1.1 $
+ Date: $Date: 2006/04/17 21:34:31 $
+ Version: $Revision: 1.3 $
Copyright (c) Insight Software Consortium. All rights reserved.
See ITKCopyright.txt or http://www.itk.org/HTML/Copyright.htm for details.
@@ -37,6 +37,7 @@ public:
typedef SmartPointer<Self> Pointer;
typedef SmartPointer<const Self> ConstPointer;
typedef typename Superclass::ErrorVectorType ErrorVectorType;
+ typedef typename Superclass::InternalVectorType InternalVectorType;
/** Run-time type information (and related methods). */
itkTypeMacro(MeanSquaredErrorFunction, FunctionBase);
@@ -47,7 +48,7 @@ public:
/** Evaluate at the specified Error position */
ScalarType Evaluate(const TVector& Errors) const;
- ErrorVectorType EvaluateDerivative(const TVector& Errors) const;
+ InternalVectorType EvaluateDerivative(const TVector& Errors) const;
protected:
diff --git a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkMeanSquaredErrorFunction.txx b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkMeanSquaredErrorFunction.txx
index 7d930df028..889e2f86fe 100644
--- a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkMeanSquaredErrorFunction.txx
+++ b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkMeanSquaredErrorFunction.txx
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkMeanSquaredErrorFunction.txx,v $
Language: C++
- Date: $Date: 2005/08/02 19:17:37 $
- Version: $Revision: 1.1 $
+ Date: $Date: 2006/04/17 19:34:45 $
+ Version: $Revision: 1.2 $
Copyright (c) Insight Software Consortium. All rights reserved.
See ITKCopyright.txt or http://www.itk.org/HTML/Copyright.htm for details.
@@ -45,19 +45,28 @@ ScalarType
MeanSquaredErrorFunction <TVector,ScalarType>
::Evaluate(const TVector& errors) const
{
- vnl_vector <ScalarType> temp(errors.GetVnlVector());
+ vnl_vector <ScalarType> temp(errors.Size());
+ for(unsigned int i=0; i<errors.Size(); i++)
+ {
+ temp[i]=errors[i];
+ }
return (temp.squared_magnitude() / temp.size());
}
/** Evaluate derivatives */
template<class TVector, class ScalarType>
typename MeanSquaredErrorFunction <TVector,ScalarType>
-::ErrorVectorType
+::InternalVectorType
MeanSquaredErrorFunction <TVector,ScalarType>
::EvaluateDerivative(const TVector& errors) const
{
- ScalarType m = static_cast<ScalarType>(2) / errors.GetVectorDimension();
- return (errors * m);
+ ScalarType m = static_cast<ScalarType>(2) / errors.Size();
+ InternalVectorType temp(errors.Size());
+ for(int i=0; i<errors.Size(); i++)
+ {
+ temp[i]=errors[i]*m;
+ }
+ return temp;
}
/** Print the object */
diff --git a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkMultilayerNeuralNetworkBase.h b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkMultilayerNeuralNetworkBase.h
index 8760724a5e..4df398c0f4 100644
--- a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkMultilayerNeuralNetworkBase.h
+++ b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkMultilayerNeuralNetworkBase.h
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkMultilayerNeuralNetworkBase.h,v $
Language: C++
- Date: $Date: 2005/11/21 18:39:31 $
- Version: $Revision: 1.2 $
+ Date: $Date: 2006/04/17 21:34:31 $
+ Version: $Revision: 1.4 $
Copyright (c) Insight Software Consortium. All rights reserved.
See ITKCopyright.txt or http://www.itk.org/HTML/Copyright.htm for details.
@@ -42,7 +42,7 @@ public:
itkNewMacro( Self );
typedef typename Superclass::ValueType ValueType;
-
+ typedef typename Superclass::NetworkOutputType NetworkOutputType;
typedef typename Superclass::LayerType LayerType;
typedef typename Superclass::WeightSetType WeightSetType;
typedef typename Superclass::WeightSetPointer WeightSetPointer;
@@ -55,17 +55,24 @@ public:
itkSetMacro(NumOfLayers, int);
itkGetConstReferenceMacro(NumOfLayers, int);
+
+ itkSetMacro(NumOfWeightSets, int);
+ itkGetConstReferenceMacro(NumOfWeightSets, int);
void AddLayer(LayerType*);
+ LayerType* GetLayer(int layer_id);
void AddWeightSet(WeightSetType*);
+ WeightSetType* GetWeightSet(int id);
void SetLearningFunction(LearningFunctionType* f);
- virtual ValueType* GenerateOutput(TVector samplevector);
-
- virtual void BackwardPropagate(TOutput errors);
+ // virtual ValueType* GenerateOutput(TVector samplevector);
+ virtual NetworkOutputType GenerateOutput(TVector samplevector);
+// virtual void BackwardPropagate(TOutput errors);
+ virtual void BackwardPropagate(NetworkOutputType errors);
+
virtual void UpdateWeights(ValueType);
void SetLearningRule(LearningFunctionType*);
@@ -83,7 +90,7 @@ protected:
LearningFunctionPointer m_LearningFunction;
ValueType m_LearningRate;
int m_NumOfLayers;
-
+ int m_NumOfWeightSets;
/** Method to print the object. */
virtual void PrintSelf( std::ostream& os, Indent indent ) const;
};
diff --git a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkMultilayerNeuralNetworkBase.txx b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkMultilayerNeuralNetworkBase.txx
index 41c70454eb..b6295179b9 100644
--- a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkMultilayerNeuralNetworkBase.txx
+++ b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkMultilayerNeuralNetworkBase.txx
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkMultilayerNeuralNetworkBase.txx,v $
Language: C++
- Date: $Date: 2005/08/03 15:47:06 $
- Version: $Revision: 1.3 $
+ Date: $Date: 2006/04/17 19:34:45 $
+ Version: $Revision: 1.4 $
Copyright (c) Insight Software Consortium. All rights reserved.
See ITKCopyright.txt or http://www.itk.org/HTML/Copyright.htm for details.
@@ -32,6 +32,7 @@ MultilayerNeuralNetworkBase<TVector,TOutput>
m_LearningFunction = DefaultLearningFunctionType::New();
m_LearningRate = 0.001;
m_NumOfLayers = 0;
+ m_NumOfWeightSets=0;
}
template<class TVector, class TOutput>
@@ -67,7 +68,17 @@ MultilayerNeuralNetworkBase<TVector,TOutput>
}
template<class TVector, class TOutput>
-typename MultilayerNeuralNetworkBase<TVector, TOutput>::ValueType*
+typename MultilayerNeuralNetworkBase<TVector, TOutput>::LayerType*
+MultilayerNeuralNetworkBase<TVector,TOutput>
+::GetLayer(int layer_id)
+{
+ return m_Layers[layer_id].GetPointer();
+}
+
+
+
+template<class TVector, class TOutput>
+typename MultilayerNeuralNetworkBase<TVector, TOutput>::NetworkOutputType
MultilayerNeuralNetworkBase<TVector,TOutput>
::GenerateOutput(TVector samplevector)
{
@@ -78,13 +89,18 @@ MultilayerNeuralNetworkBase<TVector,TOutput>
this->m_Weights[i]->ForwardPropagate(this->m_Layers[i]->GetOutputVector());
this->m_Layers[i + 1]->ForwardPropagate();
}
- return this->m_Layers[i]->GetOutputVector();
+ NetworkOutputType temp_output;
+ temp_output.SetSize(this->m_Layers[i]->GetNumberOfNodes());
+ for(unsigned int k=0; k<temp_output.Size(); k++)
+ temp_output[k]=this->m_Layers[i]->GetOutputVector()[k];
+// return this->m_Layers[i]->GetOutputVector();
+ return temp_output;
}
template<class TVector, class TOutput>
void
MultilayerNeuralNetworkBase<TVector,TOutput>
-::BackwardPropagate(TOutput errors)
+::BackwardPropagate(NetworkOutputType errors)
{
int i = this->m_Layers.size();
i--;
@@ -129,8 +145,18 @@ MultilayerNeuralNetworkBase<TVector,TOutput>
::AddWeightSet(WeightSetType* weightset)
{
m_Weights.push_back(weightset);
+ m_NumOfWeightSets++;
}
+template<class TVector, class TOutput>
+typename MultilayerNeuralNetworkBase<TVector, TOutput>::WeightSetType*
+MultilayerNeuralNetworkBase<TVector,TOutput>
+::GetWeightSet(int id)
+{
+ return m_Weights[id].GetPointer();
+}
+
+
template<class TVector, class TOutput>
void
MultilayerNeuralNetworkBase<TVector,TOutput>
diff --git a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkMultiquadricRadialBasisFunction.txx b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkMultiquadricRadialBasisFunction.txx
index 8e6f04365c..c6d9cbafe5 100644
--- a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkMultiquadricRadialBasisFunction.txx
+++ b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkMultiquadricRadialBasisFunction.txx
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkMultiquadricRadialBasisFunction.txx,v $
Language: C++
- Date: $Date: 2005/08/02 19:17:37 $
- Version: $Revision: 1.1 $
+ Date: $Date: 2006/03/19 04:37:20 $
+ Version: $Revision: 1.2 $
Copyright (c) Insight Software Consortium. All rights reserved.
See ITKCopyright.txt or http://www.itk.org/HTML/Copyright.htm for details.
@@ -45,7 +45,7 @@ ScalarType
MultiquadricRadialBasisFunction<ScalarType>
::Evaluate(const ScalarType& input) const
{
- const ScalarType val = pow((input*input)+(m_Radius*m_Radius),0.5);
+ const ScalarType val = vcl_pow((input*input)+(m_Radius*m_Radius),0.5);
return val;
}
diff --git a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkNeuralNetworkObject.h b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkNeuralNetworkObject.h
index 300348ace4..dfc1591ea7 100644
--- a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkNeuralNetworkObject.h
+++ b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkNeuralNetworkObject.h
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkNeuralNetworkObject.h,v $
Language: C++
- Date: $Date: 2005/08/02 19:17:37 $
- Version: $Revision: 1.1 $
+ Date: $Date: 2006/04/17 19:34:45 $
+ Version: $Revision: 1.2 $
Copyright (c) Insight Software Consortium. All rights reserved.
See ITKCopyright.txt or http://www.itk.org/HTML/Copyright.htm for details.
@@ -40,7 +40,8 @@ public:
typedef typename TVector::ValueType ValueType;
typedef TOutput TargetVectorType;
-
+ typedef Array<ValueType> NetworkOutputType;
+
typedef WeightSetBase<TVector, TOutput> WeightSetType;
typedef LayerBase<TVector, TOutput> LayerType;
typedef typename LayerType::Pointer LayerPointer;
@@ -48,10 +49,11 @@ public:
typedef LearningFunctionBase<LayerType, TOutput> LearningFunctionType;
typedef typename LearningFunctionType::Pointer LearningFunctionPointer;
- virtual ValueType* GenerateOutput(TVector samplevector) = 0;
-
- virtual void BackwardPropagate(TOutput errors) = 0;
+ //virtual ValueType* GenerateOutput(TVector samplevector) = 0;
+ virtual NetworkOutputType GenerateOutput(TVector samplevector)=0;
+ //virtual void BackwardPropagate(TOutput errors) = 0;
+ virtual void BackwardPropagate(NetworkOutputType errors) = 0;
virtual void UpdateWeights(ValueType) = 0;
protected:
diff --git a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkOneHiddenLayerBackPropagationNeuralNetwork.h b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkOneHiddenLayerBackPropagationNeuralNetwork.h
index 97c64661da..d65b704d75 100644
--- a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkOneHiddenLayerBackPropagationNeuralNetwork.h
+++ b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkOneHiddenLayerBackPropagationNeuralNetwork.h
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkOneHiddenLayerBackPropagationNeuralNetwork.h,v $
Language: C++
- Date: $Date: 2005/08/03 14:40:14 $
- Version: $Revision: 1.2 $
+ Date: $Date: 2006/04/17 21:34:31 $
+ Version: $Revision: 1.4 $
Copyright (c) Insight Software Consortium. All rights reserved.
See ITKCopyright.txt or http://www.itk.org/HTML/Copyright.htm for details.
@@ -50,7 +50,8 @@ public:
typedef SmartPointer<Self> Pointer;
typedef SmartPointer<const Self> ConstPointer;
typedef typename Superclass::ValueType ValueType;
-
+ typedef typename Superclass::NetworkOutputType NetworkOutputType;
+
typedef TransferFunctionBase<ValueType> TransferFunctionType;
typedef InputFunctionBase<ValueType*, ValueType> InputFunctionType;
@@ -78,7 +79,8 @@ public:
itkSetMacro(OutputLayerBias, ValueType);
itkGetConstReferenceMacro(OutputLayerBias, ValueType);
- ValueType* GenerateOutput(TVector samplevector);
+ //ValueType* GenerateOutput(TVector samplevector);
+ NetworkOutputType GenerateOutput(TVector samplevector);
void SetInputTransferFunction(TransferFunctionType* f);
void SetHiddenTransferFunction(TransferFunctionType* f);
diff --git a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkOneHiddenLayerBackPropagationNeuralNetwork.txx b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkOneHiddenLayerBackPropagationNeuralNetwork.txx
index 4c83606801..d1a25d9e88 100644
--- a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkOneHiddenLayerBackPropagationNeuralNetwork.txx
+++ b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkOneHiddenLayerBackPropagationNeuralNetwork.txx
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkOneHiddenLayerBackPropagationNeuralNetwork.txx,v $
Language: C++
- Date: $Date: 2005/08/02 19:17:37 $
- Version: $Revision: 1.1 $
+ Date: $Date: 2006/04/17 19:34:45 $
+ Version: $Revision: 1.2 $
Copyright (c) Insight Software Consortium. All rights reserved.
See ITKCopyright.txt or http://www.itk.org/HTML/Copyright.htm for details.
@@ -144,7 +144,7 @@ OneHiddenLayerBackPropagationNeuralNetwork<TVector,TOutput>
/** Generate output */
template<class TVector, class TOutput>
-typename OneHiddenLayerBackPropagationNeuralNetwork<TVector, TOutput>::ValueType*
+typename OneHiddenLayerBackPropagationNeuralNetwork<TVector, TOutput>::NetworkOutputType
OneHiddenLayerBackPropagationNeuralNetwork<TVector,TOutput>
::GenerateOutput(TVector samplevector)
{
diff --git a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkRBFBackPropagationLearningFunction.txx b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkRBFBackPropagationLearningFunction.txx
index 22034f4a21..6a061a33d3 100644
--- a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkRBFBackPropagationLearningFunction.txx
+++ b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkRBFBackPropagationLearningFunction.txx
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkRBFBackPropagationLearningFunction.txx,v $
Language: C++
- Date: $Date: 2005/08/03 15:47:06 $
- Version: $Revision: 1.3 $
+ Date: $Date: 2006/04/17 19:34:45 $
+ Version: $Revision: 1.4 $
Copyright (c) Insight Software Consortium. All rights reserved.
See ITKCopyright.txt or http://www.itk.org/HTML/Copyright.htm for details.
@@ -48,7 +48,7 @@ RBFBackPropagationLearningFunction<LayerType,TOutput>
typedef RBFLayer<InputVectorType,OutputVectorType>
RbfLayerType;
- typedef typename RbfLayerType::NodeArrayType ArrayType;
+ typedef typename RbfLayerType::InternalVectorType ArrayType;
typename LayerType::ValuePointer currentdeltavalues = inputweightset->GetTotalDeltaValues();
vnl_matrix<ValueType> DW_temp(currentdeltavalues,inputweightset->GetNumberOfOutputNodes(),
inputweightset->GetNumberOfInputNodes());
diff --git a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkRBFLayer.h b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkRBFLayer.h
index dd00b03f63..219d0bd79a 100644
--- a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkRBFLayer.h
+++ b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkRBFLayer.h
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkRBFLayer.h,v $
Language: C++
- Date: $Date: 2005/08/26 00:03:39 $
- Version: $Revision: 1.5 $
+ Date: $Date: 2006/04/17 21:34:31 $
+ Version: $Revision: 1.7 $
Copyright (c) Insight Software Consortium. All rights reserved.
See ITKCopyright.txt or http://www.itk.org/HTML/Copyright.htm for details.
@@ -45,18 +45,27 @@ public:
typedef typename Superclass::ValueType ValueType;
typedef typename Superclass::ValuePointer ValuePointer;
typedef vnl_vector<ValueType> NodeVectorType;
- typedef Array<ValueType> NodeArrayType;
+ //typedef Array<ValueType> NodeArrayType;
+ typedef typename Superclass::InternalVectorType InternalVectorType;
+
typedef typename Superclass::OutputVectorType OutputVectorType;
typedef RadialBasisFunctionBase<ValueType> RBFType;
//Distance Metric
- typedef EuclideanDistance<TVector> DistanceMetricType;
+ typedef EuclideanDistance<InternalVectorType> DistanceMetricType;
typedef typename DistanceMetricType::Pointer DistanceMetricPointer;
//Member Functions
- void SetNumberOfNodes(unsigned int);
+ void SetNumberOfNodes(unsigned int numNodes);
+ //void SetMeasurementVectorSize(unsigned int size);
+ itkGetMacro(RBF_Dim, unsigned int);
+ void SetRBF_Dim(unsigned int size);
+
+
ValueType GetInputValue(unsigned int i);
- void SetInputValue(unsigned int i, ValueType value);
+ void SetInputValue(unsigned int i,ValueType value);
+
+ itkGetMacro(LayerType, unsigned int);
ValueType GetOutputValue(int);
void SetOutputValue(int, ValueType);
@@ -78,7 +87,8 @@ public:
ValuePointer GetInputErrorVector();
void SetInputErrorValue(ValueType, int node_id);
- TVector GetCenter(int i);
+ //TVector GetCenter(int i);
+ InternalVectorType GetCenter(int i);
void SetCenter(TVector c,int i);
ValueType GetRadii(int i);
@@ -110,14 +120,15 @@ protected:
virtual void PrintSelf( std::ostream& os, Indent indent ) const;
private:
-
+
typename DistanceMetricType::Pointer m_DistanceMetric;
NodeVectorType m_NodeInputValues;
NodeVectorType m_NodeOutputValues;
NodeVectorType m_InputErrorValues;
NodeVectorType m_OutputErrorValues;
- std::vector<TVector> m_Centers; // ui....uc
- NodeArrayType m_Radii;
+ //std::vector<TVector> m_Centers; // ui....uc
+ std::vector<InternalVectorType> m_Centers; // ui....uc
+ InternalVectorType m_Radii;
int m_NumClasses;
ValueType m_Bias;
int m_RBF_Dim;
diff --git a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkRBFLayer.txx b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkRBFLayer.txx
index f543c93a1d..cd0c68d8cd 100644
--- a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkRBFLayer.txx
+++ b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkRBFLayer.txx
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkRBFLayer.txx,v $
Language: C++
- Date: $Date: 2005/08/10 19:53:43 $
- Version: $Revision: 1.4 $
+ Date: $Date: 2006/04/17 19:34:45 $
+ 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.
@@ -34,9 +34,9 @@ RBFLayer<TVector,TOutput>
typedef GaussianRadialBasisFunction<ValueType> GRBFType;
m_RBF=GRBFType::New();
m_DistanceMetric = DistanceMetricType::New();
- TVector origin;
+// TVector origin;
- m_DistanceMetric->SetMeasurementVectorSize(origin.GetVectorDimension());
+// m_DistanceMetric->SetMeasurementVectorSize(origin.Size());
m_RBF_Dim = 0;
}
@@ -54,14 +54,22 @@ RBFLayer<TVector,TOutput>
m_RBF = f;
this->Modified();
}
+template<class TVector, class TOutput>
+void
+RBFLayer<TVector,TOutput>
+::SetRBF_Dim(unsigned int dim)
+{
+ m_RBF_Dim=dim;
+ m_DistanceMetric->SetMeasurementVectorSize(m_RBF_Dim);
+}
template<class TVector, class TOutput>
void
RBFLayer<TVector,TOutput>
::SetNumberOfNodes(unsigned int c)
{
- TVector sampleinputvector;
- m_RBF_Dim= sampleinputvector.GetVectorDimension();
+ //TVector sampleinputvector;
+ //m_RBF_Dim= sampleinputvector.Size();
Superclass::SetNumberOfNodes(c);
this->m_NodeInputValues.set_size(m_RBF_Dim); //c);
this->m_NodeOutputValues.set_size(c);
@@ -70,7 +78,8 @@ RBFLayer<TVector,TOutput>
if(this->GetLayerType()!=3)
{
- TVector temp;
+ //TVector temp;
+ InternalVectorType temp(m_RBF_Dim);
for(unsigned int i=0; i<c; i++)
{
m_Centers.push_back(temp);
@@ -155,12 +164,17 @@ void
RBFLayer<TVector,TOutput>
::SetCenter(TVector c,int i)
{
- m_Centers[i]=c;
+ InternalVectorType temp(c.Size());
+ for(int j=0; j<c.Size(); j++)
+ {
+ temp[j]=c[j];
+ }
+ m_Centers[i]=temp; //c;
this->Modified();
}
template<class TVector, class TOutput>
-TVector
+typename RBFLayer<TVector,TOutput>::InternalVectorType
RBFLayer<TVector,TOutput>
::GetCenter(int i)
{
@@ -257,11 +271,14 @@ RBFLayer<TVector,TOutput>
width_increment.fill(0);
width_increment= center_increment.get_column(inputweightset->GetNumberOfInputNodes()-1);
ValueType temp_radius;
- NodeArrayType temp_center;
+ InternalVectorType temp_center;
temp_center.SetSize(m_RBF_Dim);
- TVector tempvector1;
- TVector tempvector2;
- TVector tempcenter;
+ //TVector tempvector1;
+ //TVector tempvector2;
+ //TVector tempcenter;
+ InternalVectorType tempvector1(m_RBF_Dim);
+ InternalVectorType tempvector2(m_RBF_Dim);
+ InternalVectorType tempcenter(m_RBF_Dim);
for (int i = 0; i < m_NumClasses; i++)
{
@@ -277,18 +294,21 @@ RBFLayer<TVector,TOutput>
temp_radius = m_Radii.GetElement(i);
temp_radius += width_increment[i];
m_Radii.SetElement(i,temp_radius);
- NodeArrayType array1(m_NodeInputValues.size());
+ InternalVectorType array1(m_NodeInputValues.size());
array1= m_NodeInputValues;
- tempvector1.Set_vnl_vector(m_NodeInputValues);
+ for(int j=0; j<tempvector1.size(); j++)
+ tempvector1[j]=m_NodeInputValues[j];
+
+ //tempvector1.Set_vnl_vector(m_NodeInputValues);
tempvector2=m_Centers[i];
tempcenter= m_Centers[i];
//double dt= m_DistanceMetric->Evaluate(tempvector1,tempvector2);
//std::cout<<"Euclidean in layer ="<<dt<<std::endl;
m_RBF->SetRadius(m_Radii.GetElement(i));
- NodeArrayType temp_array(m_RBF_Dim);
- TVector temp_vector= m_Centers[i];
+ InternalVectorType temp_array(m_RBF_Dim);
+ NodeVectorType temp_vector= m_Centers[i];
for(int ii=0; ii<m_RBF_Dim; ii++)
temp_array.SetElement(ii,temp_vector[ii]);
m_RBF->SetCenter(temp_array);
@@ -327,7 +347,11 @@ void
RBFLayer<TVector,TOutput>
::SetOutputErrorValues(TOutput errors)
{
- m_OutputErrorValues = errors.GetVnlVector();
+
+ for(unsigned int i=0; i<errors.Size(); i++)
+ m_OutputErrorValues[i] = errors[i];
+
+ this->Modified();
}
template<class TVector, class TOutput>
@@ -366,9 +390,13 @@ RBFLayer<TVector,TOutput>
vnl_vector<ValueType> deltaww;
deltaww.set_size(cols);
deltaww.fill(0);
+ /*
TVector tempvector1;
- TVector tempvector2;
-
+ TVector tempvector2;*/
+
+ InternalVectorType tempvector1(m_RBF_Dim);
+ InternalVectorType tempvector2(m_RBF_Dim);
+
for(int c1=0; c1<rows; c1++)
{
for(int c2=0; c2<cols; c2++)
@@ -383,12 +411,14 @@ RBFLayer<TVector,TOutput>
}
//compute gradient for centers
- NodeArrayType array1(m_NodeInputValues.size());
+ InternalVectorType array1(m_NodeInputValues.size());
array1= m_NodeInputValues;
vnl_matrix<ValueType> DW_temp(inputweightset->GetNumberOfOutputNodes(),
inputweightset->GetNumberOfInputNodes());
DW_temp.fill(0.0);
- tempvector1.Set_vnl_vector(array1);
+
+ for(int k=0; k<array1.Size(); k++)
+ tempvector1[k]=array1[k];
for(int k=0; k<num_nodes; k++)
{
@@ -397,8 +427,8 @@ RBFLayer<TVector,TOutput>
tempvector2=m_Centers[k];
double dist=m_DistanceMetric->Evaluate(tempvector1,tempvector2);
m_RBF->SetRadius(m_Radii.GetElement(k));
- TVector temp_vector= m_Centers[k];
- NodeArrayType temp_array(m_RBF_Dim);
+ NodeVectorType temp_vector= m_Centers[k];
+ InternalVectorType temp_array(m_RBF_Dim);
for(int ii=0; ii<m_RBF_Dim; ii++)
temp_array.SetElement(ii,temp_vector[ii]);
m_RBF->SetCenter(temp_array);
diff --git a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkRBFNetwork.h b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkRBFNetwork.h
index 565db5a40c..7c61cd8d04 100644
--- a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkRBFNetwork.h
+++ b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkRBFNetwork.h
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkRBFNetwork.h,v $
Language: C++
- Date: $Date: 2005/08/03 14:40:14 $
- Version: $Revision: 1.2 $
+ Date: $Date: 2006/04/18 11:23:29 $
+ Version: $Revision: 1.4 $
Copyright (c) Insight Software Consortium. All rights reserved.
See ITKCopyright.txt or http://www.itk.org/HTML/Copyright.htm for details.
@@ -49,11 +49,10 @@ public:
typedef SmartPointer<const Self> ConstPointer;
typedef typename Superclass::ValueType ValueType;
typedef Array<ValueType> ArrayType;
-
typedef TransferFunctionBase<ValueType> TransferFunctionType;
typedef RadialBasisFunctionBase<ValueType> RBFType;
typedef InputFunctionBase<ValueType*, ValueType> InputFunctionType;
- typedef EuclideanDistance<TVector> DistanceMetricType;
+ typedef EuclideanDistance<ArrayType> DistanceMetricType;
typename InputFunctionType::Pointer InputFunction;
typename DistanceMetricType::Pointer DistanceMetric;
@@ -62,6 +61,8 @@ public:
typename RBFType::Pointer HiddenTransferFunction;
typename TransferFunctionType::Pointer OutputTransferFunction;
+ typedef typename Superclass::NetworkOutputType NetworkOutputType;
+
/* Method for creation through the object factory. */
itkTypeMacro(RBFNetwork,
MultilayerNeuralNetworkBase);
@@ -89,7 +90,8 @@ public:
itkSetMacro(Classes, int);
itkGetConstReferenceMacro(Classes,int);
- ValueType* GenerateOutput(TVector samplevector);
+ // ValueType* GenerateOutput(TVector samplevector);
+ virtual NetworkOutputType GenerateOutput(TVector samplevector);
void SetInputTransferFunction(TransferFunctionType* f);
void SetDistanceMetric(DistanceMetricType* f);
@@ -99,6 +101,9 @@ public:
void SetInputFunction(InputFunctionType* f);
void InitializeWeights();
+ void SetCenter(TVector c);
+ void SetRadius(ValueType r);
+
protected:
RBFNetwork();
@@ -115,6 +120,8 @@ private:
int m_Classes;
ValueType m_HiddenLayerBias;
ValueType m_OutputLayerBias;
+ std::vector<TVector> m_Centers; // ui....uc
+ std::vector<double> m_Radii;
};
} // end namespace Statistics
diff --git a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkRBFNetwork.txx b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkRBFNetwork.txx
index f4c638ebbc..97ccb8ccbe 100644
--- a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkRBFNetwork.txx
+++ b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkRBFNetwork.txx
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkRBFNetwork.txx,v $
Language: C++
- Date: $Date: 2005/08/02 19:17:37 $
- Version: $Revision: 1.1 $
+ Date: $Date: 2006/04/17 19:34:45 $
+ Version: $Revision: 1.2 $
Copyright (c) Insight Software Consortium. All rights reserved.
See ITKCopyright.txt or http://www.itk.org/HTML/Copyright.htm for details.
@@ -81,10 +81,12 @@ RBFNetwork<TVector,TOutput>
typename rbflayertype::Pointer hiddenlayer = rbflayertype::New();
+ hiddenlayer->SetRBF_Dim(m_NumOfInputNodes);
hiddenlayer->SetLayerType(2);
hiddenlayer->SetNumClasses(m_Classes);
hiddenlayer->SetNumberOfNodes(m_NumOfHiddenNodes);
+
typename bplayertype::Pointer outputlayer = bplayertype::New();
hiddenlayer->SetNumClasses(m_Classes);
outputlayer->SetLayerType(3);
@@ -124,7 +126,7 @@ RBFNetwork<TVector,TOutput>
Superclass::AddWeightSet(IW);
Superclass::AddWeightSet(HW);
-
+ /*
TVector temp1;
TVector temp2;
temp1[0]=110;
@@ -142,7 +144,13 @@ RBFNetwork<TVector,TOutput>
hiddenlayer->SetRadii(2*width,0);
hiddenlayer->SetRadii(2*width,1);
-
+ */
+ for(int j=0; j<m_Centers.size(); j++)
+ {
+ hiddenlayer->SetCenter(m_Centers[j],j);
+ hiddenlayer->SetRadii(m_Radii[j],j);
+ }
+
}
template<class TVector, class TOutput>
@@ -186,13 +194,29 @@ RBFNetwork<TVector,TOutput>
}
template<class TVector, class TOutput>
-typename RBFNetwork<TVector, TOutput>::ValueType*
+typename RBFNetwork<TVector, TOutput>::NetworkOutputType
RBFNetwork<TVector,TOutput>
::GenerateOutput(TVector samplevector)
{
return Superclass::GenerateOutput(samplevector);
}
+template<class TVector, class TOutput>
+void
+RBFNetwork<TVector,TOutput>
+::SetCenter(TVector c)
+{
+ m_Centers.push_back(c);
+}
+
+template<class TVector, class TOutput>
+void
+RBFNetwork<TVector,TOutput>
+::SetRadius(ValueType r)
+{
+ m_Radii.push_back(r);
+}
+
/** Print the object */
template<class TVector, class TOutput>
void
diff --git a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkSigmoidTransferFunction.txx b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkSigmoidTransferFunction.txx
index 50457fa139..0307bb8325 100644
--- a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkSigmoidTransferFunction.txx
+++ b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkSigmoidTransferFunction.txx
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkSigmoidTransferFunction.txx,v $
Language: C++
- Date: $Date: 2005/08/02 19:17:37 $
- Version: $Revision: 1.1 $
+ Date: $Date: 2006/03/19 04:37:20 $
+ Version: $Revision: 1.2 $
Copyright (c) Insight Software Consortium. All rights reserved.
See ITKCopyright.txt or http://www.itk.org/HTML/Copyright.htm for details.
@@ -49,7 +49,7 @@ SigmoidTransferFunction<ScalarType>
::Evaluate(const ScalarType& input) const
{
const ScalarType x = (static_cast<ScalarType>(input) - m_Beta) / m_Alpha;
- const ScalarType e = 1.0 / (1.0 + exp(-x));
+ const ScalarType e = 1.0 / (1.0 + vcl_exp(-x));
const ScalarType v = (m_OutputMaximum - m_OutputMinimum) * e
+ m_OutputMinimum;
return v;
diff --git a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkSquaredDifferenceErrorFunction.h b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkSquaredDifferenceErrorFunction.h
index 9ac9232e41..b0a7946761 100644
--- a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkSquaredDifferenceErrorFunction.h
+++ b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkSquaredDifferenceErrorFunction.h
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkSquaredDifferenceErrorFunction.h,v $
Language: C++
- Date: $Date: 2005/08/02 19:17:37 $
- Version: $Revision: 1.1 $
+ Date: $Date: 2006/04/17 21:34:31 $
+ Version: $Revision: 1.3 $
Copyright (c) Insight Software Consortium. All rights reserved.
See ITKCopyright.txt or http://www.itk.org/HTML/Copyright.htm for details.
@@ -37,6 +37,7 @@ public:
typedef SmartPointer<Self> Pointer;
typedef SmartPointer<const Self> ConstPointer;
typedef typename Superclass::ErrorVectorType ErrorVectorType;
+ typedef typename Superclass::InternalVectorType InternalVectorType;
/** Run-time type information (and related methods). */
itkTypeMacro(SquaredDifferenceErrorFunction, FunctionBase);
@@ -47,7 +48,7 @@ public:
/** Evaluate at the specified Error position */
ScalarType Evaluate(const TVector& Errors) const;
- ErrorVectorType EvaluateDerivative(const TVector& Errors) const;
+ InternalVectorType EvaluateDerivative(const TVector& Errors) const;
protected:
diff --git a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkSquaredDifferenceErrorFunction.txx b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkSquaredDifferenceErrorFunction.txx
index 8f8d3b9995..027e05da74 100644
--- a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkSquaredDifferenceErrorFunction.txx
+++ b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkSquaredDifferenceErrorFunction.txx
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkSquaredDifferenceErrorFunction.txx,v $
Language: C++
- Date: $Date: 2005/08/03 12:06:12 $
- Version: $Revision: 1.2 $
+ Date: $Date: 2006/04/17 19:34:45 $
+ Version: $Revision: 1.3 $
Copyright (c) Insight Software Consortium. All rights reserved.
See ITKCopyright.txt or http://www.itk.org/HTML/Copyright.htm for details.
@@ -45,19 +45,27 @@ ScalarType
SquaredDifferenceErrorFunction <TVector,ScalarType>
::Evaluate(const TVector& errors) const
{
- vnl_vector <ScalarType> temp(errors.GetVnlVector());
+ vnl_vector <ScalarType> temp;
+ temp.set_size(errors.Size());
+ for(unsigned int i=0; i<errors.Size(); i++)
+ temp[i]=errors[i];
+
+ //temp.(errors.GetVnlVector());
+
return (temp.squared_magnitude() / 2);
}
/** Evaluate derivatives */
template<class TVector, class ScalarType>
typename SquaredDifferenceErrorFunction <TVector,ScalarType>
-::ErrorVectorType
+::InternalVectorType
SquaredDifferenceErrorFunction <TVector,ScalarType>
::EvaluateDerivative(const TVector& errors) const
{
- TVector diff;
- for(unsigned int i=0; i<errors.GetVectorDimension(); i++)
+ //TVector diff;
+ InternalVectorType diff;
+ diff.SetSize(errors.Size());
+ for(unsigned int i=0; i<errors.Size(); i++)
{
if ((-0.1 < errors[i]) && (errors[i] < 0.1))
{
diff --git a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkSymmetricSigmoidTransferFunction.txx b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkSymmetricSigmoidTransferFunction.txx
index 5dadf8d839..d938219f5b 100644
--- a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkSymmetricSigmoidTransferFunction.txx
+++ b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkSymmetricSigmoidTransferFunction.txx
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkSymmetricSigmoidTransferFunction.txx,v $
Language: C++
- Date: $Date: 2005/08/03 17:30:12 $
- Version: $Revision: 1.2 $
+ Date: $Date: 2006/03/19 04:37:20 $
+ Version: $Revision: 1.3 $
Copyright (c) Insight Software Consortium. All rights reserved.
See ITKCopyright.txt or http://www.itk.org/HTML/Copyright.htm for details.
@@ -54,7 +54,7 @@ SymmetricSigmoidTransferFunction<ScalarType>
}
else
{
- val= (ScalarType)1.0/(1.0+exp(-input))-0.5;
+ val= (ScalarType)1.0/(1.0+vcl_exp(-input))-0.5;
}
return val;
}
diff --git a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkTanHTransferFunction.txx b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkTanHTransferFunction.txx
index 47ae4a66b0..e6bec9fec8 100644
--- a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkTanHTransferFunction.txx
+++ b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkTanHTransferFunction.txx
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkTanHTransferFunction.txx,v $
Language: C++
- Date: $Date: 2005/08/02 19:17:37 $
- Version: $Revision: 1.1 $
+ Date: $Date: 2006/03/19 04:37:20 $
+ Version: $Revision: 1.2 $
Copyright (c) Insight Software Consortium. All rights reserved.
See ITKCopyright.txt or http://www.itk.org/HTML/Copyright.htm for details.
@@ -42,8 +42,8 @@ ScalarType
TanHTransferFunction<ScalarType>
::Evaluate(const ScalarType& input) const
{
- ScalarType x = exp(input);
- ScalarType y = exp(-input);
+ ScalarType x = vcl_exp(input);
+ ScalarType y = vcl_exp(-input);
return static_cast<ScalarType>((float) (x - y) / (x + y));
}
@@ -54,7 +54,7 @@ TanHTransferFunction<ScalarType>
::EvaluateDerivative(const ScalarType& input) const
{
ScalarType f = Evaluate(input);
- return 1 - pow(f, 2);
+ return 1 - vcl_pow(f, 2);
}
/** Print the object */
diff --git a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkTanSigmoidTransferFunction.txx b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkTanSigmoidTransferFunction.txx
index b81c7d74dc..70a1e3d046 100644
--- a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkTanSigmoidTransferFunction.txx
+++ b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkTanSigmoidTransferFunction.txx
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkTanSigmoidTransferFunction.txx,v $
Language: C++
- Date: $Date: 2005/08/02 19:17:37 $
- Version: $Revision: 1.1 $
+ Date: $Date: 2006/03/19 04:37:20 $
+ Version: $Revision: 1.2 $
Copyright (c) Insight Software Consortium. All rights reserved.
See ITKCopyright.txt or http://www.itk.org/HTML/Copyright.htm for details.
@@ -43,7 +43,7 @@ TanSigmoidTransferFunction<ScalarType>
::Evaluate(const ScalarType& input) const
{
return static_cast<ScalarType>((2
- / (1 + exp(-2 * static_cast<ScalarType>(input))))
+ / (1 + vcl_exp(-2 * static_cast<ScalarType>(input))))
- 1);
}
diff --git a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkTrainingFunctionBase.h b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkTrainingFunctionBase.h
index 5556548fac..3d5413812c 100644
--- a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkTrainingFunctionBase.h
+++ b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkTrainingFunctionBase.h
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkTrainingFunctionBase.h,v $
Language: C++
- Date: $Date: 2005/08/10 18:28:38 $
- Version: $Revision: 1.4 $
+ Date: $Date: 2006/04/18 11:23:29 $
+ Version: $Revision: 1.6 $
Copyright (c) Insight Software Consortium. All rights reserved.
See ITKCopyright.txt or http://www.itk.org/HTML/Copyright.htm for details.
@@ -43,15 +43,17 @@ public:
/** Method for creation through the object factory. */
itkNewMacro(Self);
+ typedef ScalarType ValueType;
typedef typename TSample::MeasurementVectorType VectorType;
typedef typename TOutput::MeasurementVectorType OutputVectorType;
+ typedef Array<ValueType> InternalVectorType;
typedef std::vector<VectorType> InputSampleVectorType;
typedef std::vector<OutputVectorType> OutputSampleVectorType;
- typedef ScalarType ValueType;
typedef NeuralNetworkObject<VectorType, OutputVectorType> NetworkType;
- typedef ErrorFunctionBase<OutputVectorType, ScalarType> PerformanceFunctionType;
- typedef SquaredDifferenceErrorFunction<OutputVectorType, ScalarType> DefaultPerformanceType;
+ typedef ErrorFunctionBase<InternalVectorType, ScalarType> PerformanceFunctionType;
+ typedef SquaredDifferenceErrorFunction<InternalVectorType, ScalarType> DefaultPerformanceType;
+ //typedef MeanSquaredErrorFunction<InternalVectorType, ScalarType> DefaultPerformanceType;
void SetTrainingSamples(TSample* samples);
void SetTargetValues(TOutput* targets);
@@ -85,6 +87,7 @@ public:
targetconverter(typename TOutput::MeasurementVectorType v)
{
OutputVectorType temp;
+
for (unsigned int i = 0; i < v.Size(); i++)
{
temp[i] = static_cast<ScalarType>(v[i]) ;
diff --git a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkTrainingFunctionBase.txx b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkTrainingFunctionBase.txx
index fdb055ea8d..cb13e0ffc2 100644
--- a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkTrainingFunctionBase.txx
+++ b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkTrainingFunctionBase.txx
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkTrainingFunctionBase.txx,v $
Language: C++
- Date: $Date: 2005/08/02 19:17:37 $
- Version: $Revision: 1.1 $
+ Date: $Date: 2006/04/17 19:34:45 $
+ Version: $Revision: 1.2 $
Copyright (c) Insight Software Consortium. All rights reserved.
See ITKCopyright.txt or http://www.itk.org/HTML/Copyright.htm for details.
@@ -45,7 +45,9 @@ void TrainingFunctionBase<TSample,TOutput,ScalarType>
typename TSample::ConstIterator iter = samples->Begin();
while (iter != samples->End())
{
- m_InputSamples.push_back(defaultconverter(iter.GetMeasurementVector()));
+ //m_InputSamples.push_back(defaultconverter(iter.GetMeasurementVector()));
+ m_InputSamples.push_back(iter.GetMeasurementVector());
+
++iter;
}
}
@@ -57,7 +59,8 @@ void TrainingFunctionBase<TSample,TOutput,ScalarType>
typename TOutput::ConstIterator iter = targets->Begin();
while (iter != targets->End())
{
- m_Targets.push_back(targetconverter(iter.GetMeasurementVector()));
+ //m_Targets.push_back(targetconverter(iter.GetMeasurementVector()));
+ m_Targets.push_back(iter.GetMeasurementVector());
++iter;
}
std::cout << "Num of Sample Targets converted= " << m_Targets.size()
diff --git a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkTwoHiddenLayerBackPropagationNeuralNetwork.h b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkTwoHiddenLayerBackPropagationNeuralNetwork.h
index e33af2593a..f56a8f9dde 100644
--- a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkTwoHiddenLayerBackPropagationNeuralNetwork.h
+++ b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkTwoHiddenLayerBackPropagationNeuralNetwork.h
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkTwoHiddenLayerBackPropagationNeuralNetwork.h,v $
Language: C++
- Date: $Date: 2005/08/02 19:17:37 $
- Version: $Revision: 1.1 $
+ Date: $Date: 2006/04/18 11:23:29 $
+ Version: $Revision: 1.3 $
Copyright (c) Insight Software Consortium. All rights reserved.
See ITKCopyright.txt or http://www.itk.org/HTML/Copyright.htm for details.
@@ -48,6 +48,8 @@ public:
typedef SmartPointer<Self> Pointer;
typedef SmartPointer<const Self> ConstPointer;
+ typedef typename Superclass::NetworkOutputType NetworkOutputType;
+
/* Method for creation through the object factory. */
itkTypeMacro(TwoHiddenLayerBackPropagationNeuralNetwork,
MultilayerNeuralNetworkBase);
@@ -71,7 +73,8 @@ public:
itkSetMacro(NumOfOutputNodes, int);
itkGetConstReferenceMacro(NumOfOutputNodes, int);
- ValueType* GenerateOutput(TVector samplevector);
+ //ValueType* GenerateOutput(TVector samplevector);
+ virtual NetworkOutputType GenerateOutput(TVector samplevector);
protected:
diff --git a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkTwoHiddenLayerBackPropagationNeuralNetwork.txx b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkTwoHiddenLayerBackPropagationNeuralNetwork.txx
index 4ae782fd29..63f78ec70c 100644
--- a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkTwoHiddenLayerBackPropagationNeuralNetwork.txx
+++ b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkTwoHiddenLayerBackPropagationNeuralNetwork.txx
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkTwoHiddenLayerBackPropagationNeuralNetwork.txx,v $
Language: C++
- Date: $Date: 2005/08/02 19:17:37 $
- Version: $Revision: 1.1 $
+ Date: $Date: 2006/04/17 19:34:45 $
+ Version: $Revision: 1.2 $
Copyright (c) Insight Software Consortium. All rights reserved.
See ITKCopyright.txt or http://www.itk.org/HTML/Copyright.htm for details.
@@ -123,7 +123,7 @@ TwoHiddenLayerBackPropagationNeuralNetwork<TVector,TOutput>
}
template<class TVector, class TOutput>
-typename TwoHiddenLayerBackPropagationNeuralNetwork<TVector, TOutput>::ValueType*
+typename TwoHiddenLayerBackPropagationNeuralNetwork<TVector, TOutput>::NetworkOutputType
TwoHiddenLayerBackPropagationNeuralNetwork<TVector,TOutput>
::GenerateOutput(TVector samplevector)
{
diff --git a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkWeightSetBase.h b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkWeightSetBase.h
index 71b8c5232a..5e08c1a722 100644
--- a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkWeightSetBase.h
+++ b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkWeightSetBase.h
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkWeightSetBase.h,v $
Language: C++
- Date: $Date: 2005/08/10 20:36:01 $
- Version: $Revision: 1.2 $
+ Date: $Date: 2006/04/17 19:34:45 $
+ Version: $Revision: 1.3 $
Copyright (c) Insight Software Consortium. All rights reserved.
See ITKCopyright.txt or http://www.itk.org/HTML/Copyright.htm for details.
@@ -59,11 +59,11 @@ public:
void SetConnectivityMatrix(vnl_matrix < int>);
- void SetNumberOfInputNodes(int n);
- int GetNumberOfInputNodes();
+ void SetNumberOfInputNodes(unsigned int n);
+ unsigned int GetNumberOfInputNodes();
- void SetNumberOfOutputNodes(int n);
- int GetNumberOfOutputNodes();
+ void SetNumberOfOutputNodes(unsigned int n);
+ unsigned int GetNumberOfOutputNodes();
void SetRange(ValueType Range);
@@ -102,6 +102,7 @@ public:
ValuePointer GetPrevDeltaBValues();
ValuePointer GetWeightValues();
+
void SetWeightValues(ValuePointer weights);
@@ -125,6 +126,15 @@ public:
void InitializeWeights();
+ itkSetMacro(WeightSetId,int);
+ itkGetMacro(WeightSetId,int);
+
+ itkSetMacro(InputLayerId,int);
+ itkGetMacro(InputLayerId,int);
+
+ itkSetMacro(OutputLayerId,int);
+ itkGetMacro(OutputLayerId,int);
+
protected:
WeightSetBase();
@@ -134,8 +144,8 @@ protected:
virtual void PrintSelf( std::ostream& os, Indent indent ) const;
typename RandomVariateGeneratorType::Pointer m_RandomGenerator;
- int m_NumberOfInputNodes;
- int m_NumberOfOutputNodes;
+ unsigned int m_NumberOfInputNodes;
+ unsigned int m_NumberOfOutputNodes;
vnl_matrix<ValueType> m_OutputValues;
vnl_matrix<ValueType> m_InputErrorValues;
@@ -175,6 +185,10 @@ protected:
bool m_FirstPass;
bool m_SecondPass;
ValueType m_Range;
+ int m_InputLayerId;
+ int m_OutputLayerId;
+ int m_WeightSetId;
+
}; //class
} // end namespace Statistics
diff --git a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkWeightSetBase.txx b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkWeightSetBase.txx
index 82b1ae9fe5..3c0baa25bb 100644
--- a/Utilities/ITK/Code/Numerics/NeuralNetworks/itkWeightSetBase.txx
+++ b/Utilities/ITK/Code/Numerics/NeuralNetworks/itkWeightSetBase.txx
@@ -43,14 +43,14 @@ WeightSetBase<TVector,TOutput>
template<class TVector, class TOutput>
void
WeightSetBase<TVector,TOutput>
-::SetNumberOfInputNodes(int n)
+::SetNumberOfInputNodes(unsigned int n)
{
m_NumberOfInputNodes = n + 1; //including bias
this->Modified();
}
template<class TVector, class TOutput>
-int
+unsigned int
WeightSetBase<TVector,TOutput>
::GetNumberOfInputNodes()
{
@@ -60,14 +60,14 @@ WeightSetBase<TVector,TOutput>
template<class TVector, class TOutput>
void
WeightSetBase<TVector,TOutput>
-::SetNumberOfOutputNodes(int n)
+::SetNumberOfOutputNodes(unsigned int n)
{
m_NumberOfOutputNodes = n;
this->Modified();
}
template<class TVector, class TOutput>
-int
+unsigned int
WeightSetBase<TVector,TOutput>
::GetNumberOfOutputNodes()
{
@@ -130,15 +130,18 @@ void
WeightSetBase<TVector,TOutput>
::InitializeWeights()
{
- int num_rows = m_WeightMatrix.rows();
- int num_cols = m_WeightMatrix.cols();
+ unsigned int num_rows = m_WeightMatrix.rows();
+ unsigned int num_cols = m_WeightMatrix.cols();
+ std::cout<<num_rows <<" "<<num_cols<<std::endl;
+ std::cout<<"conectivity matrix size = "<<m_ConnectivityMatrix.rows()<<" "
+ << m_ConnectivityMatrix.cols()<<std::endl;
- for (int i = 0; i < num_rows; i++)
+ for (unsigned int i = 0; i < num_rows; i++)
{
- for (int j = 0; j < num_cols; j++)
+ for (unsigned int j = 0; j < num_cols; j++)
{
if(m_ConnectivityMatrix[i][j]==1)
- m_WeightMatrix(i, j) = RandomWeightValue(-m_Range,m_Range);
+ m_WeightMatrix(i, j) = RandomWeightValue(-1*m_Range,m_Range);
else
m_WeightMatrix(i, j) = 0;
}
@@ -208,8 +211,8 @@ WeightSetBase<TVector,TOutput>
W_temp.copy_in(w);
m_WeightMatrix = W_temp;
- ValueType v=0.0;
- m_WeightMatrix.set_column( m_NumberOfInputNodes-1,v);
+ // ValueType v=0.0;
+ // m_WeightMatrix.set_column( m_NumberOfInputNodes-1,v);
this->Modified();
}
diff --git a/Utilities/ITK/Code/Numerics/Statistics/CMakeLists.txt b/Utilities/ITK/Code/Numerics/Statistics/CMakeLists.txt
index dbe2420ee9..bccce71715 100644
--- a/Utilities/ITK/Code/Numerics/Statistics/CMakeLists.txt
+++ b/Utilities/ITK/Code/Numerics/Statistics/CMakeLists.txt
@@ -7,5 +7,8 @@ ADD_LIBRARY(ITKStatistics
TARGET_LINK_LIBRARIES(ITKStatistics ITKCommon)
-INSTALL_TARGETS(/lib/otb ITKStatistics)
-INSTALL_FILES(/include/otb/Utilities/ITK/Numerics/Statistics "(\\.h|\\.txx)$")
+IF(ITK_LIBRARY_PROPERTIES)
+ SET_TARGET_PROPERTIES(ITKStatistics PROPERTIES ${ITK_LIBRARY_PROPERTIES})
+ENDIF(ITK_LIBRARY_PROPERTIES)
+INSTALL_TARGETS(/lib/InsightToolkit ITKStatistics)
+INSTALL_FILES(/include/InsightToolkit/Numerics/Statistics "(\\.h|\\.txx)$")
diff --git a/Utilities/ITK/Code/Numerics/Statistics/itkEuclideanDistance.txx b/Utilities/ITK/Code/Numerics/Statistics/itkEuclideanDistance.txx
index b10fc43008..4075948312 100644
--- a/Utilities/ITK/Code/Numerics/Statistics/itkEuclideanDistance.txx
+++ b/Utilities/ITK/Code/Numerics/Statistics/itkEuclideanDistance.txx
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkEuclideanDistance.txx,v $
Language: C++
- Date: $Date: 2005/07/26 16:03:09 $
- Version: $Revision: 1.11 $
+ Date: $Date: 2006/03/19 04:37:20 $
+ Version: $Revision: 1.12 $
Copyright (c) Insight Software Consortium. All rights reserved.
See ITKCopyright.txt or http://www.itk.org/HTML/Copyright.htm for details.
@@ -40,7 +40,7 @@ EuclideanDistance< TVector >
distance += temp * temp ;
}
- return sqrt(distance) ;
+ return vcl_sqrt(distance) ;
}
@@ -66,7 +66,7 @@ EuclideanDistance< TVector >
distance += temp * temp ;
}
- return sqrt(distance) ;
+ return vcl_sqrt(distance) ;
}
template< class TVector >
@@ -75,7 +75,7 @@ EuclideanDistance< TVector >
::Evaluate(const ValueType &a, const ValueType &b) const
{
double temp = a - b ;
- return sqrt(temp * temp) ;
+ return vcl_sqrt(temp * temp) ;
}
template< class TVector >
diff --git a/Utilities/ITK/Code/Numerics/Statistics/itkExpectationMaximizationMixtureModelEstimator.txx b/Utilities/ITK/Code/Numerics/Statistics/itkExpectationMaximizationMixtureModelEstimator.txx
index dded9149ac..d84f126e16 100644
--- a/Utilities/ITK/Code/Numerics/Statistics/itkExpectationMaximizationMixtureModelEstimator.txx
+++ b/Utilities/ITK/Code/Numerics/Statistics/itkExpectationMaximizationMixtureModelEstimator.txx
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkExpectationMaximizationMixtureModelEstimator.txx,v $
Language: C++
- Date: $Date: 2005/05/06 18:53:40 $
- Version: $Revision: 1.15 $
+ Date: $Date: 2006/03/19 04:37:20 $
+ Version: $Revision: 1.16 $
Copyright (c) Insight Software Consortium. All rights reserved.
See ITKCopyright.txt or http://www.itk.org/HTML/Copyright.htm for details.
@@ -223,7 +223,7 @@ ExpectationMaximizationMixtureModelEstimator< TSample >
for (componentIndex = 0 ; componentIndex < m_ComponentVector.size() ;
componentIndex++)
{
- logProportion = log(m_Proportions[componentIndex]) ;
+ logProportion = vcl_log(m_Proportions[componentIndex]) ;
for (measurementVectorIndex = 0 ; measurementVectorIndex < size ;
measurementVectorIndex++)
{
@@ -231,7 +231,7 @@ ExpectationMaximizationMixtureModelEstimator< TSample >
GetWeight(measurementVectorIndex) ;
sum +=
temp * ( logProportion +
- log( m_ComponentVector[componentIndex]->
+ vcl_log(m_ComponentVector[componentIndex]->
GetWeight(measurementVectorIndex) ) ) ;
}
}
diff --git a/Utilities/ITK/Code/Numerics/Statistics/itkGaussianDensityFunction.txx b/Utilities/ITK/Code/Numerics/Statistics/itkGaussianDensityFunction.txx
index fa6f1ab5ee..9597c33ed8 100644
--- a/Utilities/ITK/Code/Numerics/Statistics/itkGaussianDensityFunction.txx
+++ b/Utilities/ITK/Code/Numerics/Statistics/itkGaussianDensityFunction.txx
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkGaussianDensityFunction.txx,v $
Language: C++
- Date: $Date: 2005/07/26 15:54:55 $
- Version: $Revision: 1.21 $
+ Date: $Date: 2006/03/19 04:37:20 $
+ Version: $Revision: 1.22 $
Copyright (c) Insight Software Consortium. All rights reserved.
See ITKCopyright.txt or http://www.itk.org/HTML/Copyright.htm for details.
@@ -97,7 +97,7 @@ GaussianDensityFunction< TMeasurementVector >
// calculate coefficient C of multivariate gaussian
m_PreFactor = 1.0 / (sqrt(det) *
- pow(sqrt(2.0 * vnl_math::pi), double(this->GetMeasurementVectorSize()))) ;
+ vcl_pow(sqrt(2.0 * vnl_math::pi), double(this->GetMeasurementVectorSize()))) ;
}
}
@@ -152,7 +152,7 @@ GaussianDensityFunction< TMeasurementVector >
temp += tempVector2[i] * tempVector[i] ;
}
- return m_PreFactor * exp( -0.5 * temp ) ;
+ return m_PreFactor * vcl_exp(-0.5 * temp ) ;
}
else
{
diff --git a/Utilities/ITK/Code/Numerics/Statistics/itkGaussianGoodnessOfFitComponent.txx b/Utilities/ITK/Code/Numerics/Statistics/itkGaussianGoodnessOfFitComponent.txx
index d9f3125290..866efbcf4f 100644
--- a/Utilities/ITK/Code/Numerics/Statistics/itkGaussianGoodnessOfFitComponent.txx
+++ b/Utilities/ITK/Code/Numerics/Statistics/itkGaussianGoodnessOfFitComponent.txx
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkGaussianGoodnessOfFitComponent.txx,v $
Language: C++
- Date: $Date: 2005/07/26 15:54:55 $
- Version: $Revision: 1.10 $
+ Date: $Date: 2006/03/19 04:37:20 $
+ Version: $Revision: 1.11 $
Copyright (c) Insight Software Consortium. All rights reserved.
See ITKCopyright.txt or http://www.itk.org/HTML/Copyright.htm for details.
@@ -226,7 +226,7 @@ GaussianGoodnessOfFitComponent< TInputSample >
double mean = 0.0 ;
double standardDeviation = 1.0 ;
- double nx = fabs( x - mean ) / standardDeviation ;
+ double nx = vcl_fabs(x - mean ) / standardDeviation ;
double t = 1 / (1 + 0.2316419 * nx);
double tt = t*t;
@@ -234,7 +234,7 @@ GaussianGoodnessOfFitComponent< TInputSample >
double tttt = ttt*t;
double ttttt = tttt*t;
- double z = exp(-nx*nx/2) / sqrt(2*vnl_math::pi);
+ double z = vcl_exp(-nx*nx/2) / vcl_sqrt(2*vnl_math::pi);
if ( x > mean )
{
diff --git a/Utilities/ITK/Code/Numerics/Statistics/itkGaussianMixtureModelComponent.txx b/Utilities/ITK/Code/Numerics/Statistics/itkGaussianMixtureModelComponent.txx
index dfab7aee82..b8f82d9ab4 100644
--- a/Utilities/ITK/Code/Numerics/Statistics/itkGaussianMixtureModelComponent.txx
+++ b/Utilities/ITK/Code/Numerics/Statistics/itkGaussianMixtureModelComponent.txx
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkGaussianMixtureModelComponent.txx,v $
Language: C++
-Date: $Date: 2005/07/26 15:54:55 $
-Version: $Revision: 1.15 $
+Date: $Date: 2006/03/19 04:37:20 $
+Version: $Revision: 1.16 $
Copyright (c) Insight Software Consortium. All rights reserved.
See ITKCopyright.txt or http://www.itk.org/HTML/Copyright.htm for details.
@@ -149,7 +149,7 @@ GaussianMixtureModelComponent< TSample >
}
}
- changes = sqrt(changes) ;
+ changes = vcl_sqrt(changes) ;
return changes ;
}
@@ -177,7 +177,7 @@ GaussianMixtureModelComponent< TSample >
{
temp = m_Mean[i] - meanEstimate[i] ;
changes = temp * temp ;
- changes = sqrt(changes) ;
+ changes = vcl_sqrt(changes) ;
if ( changes > this->GetMinimalParametersChange() )
{
changed = true ;
@@ -210,7 +210,7 @@ GaussianMixtureModelComponent< TSample >
temp = m_Covariance.GetVnlMatrix().get(i, j) -
covEstimate.GetVnlMatrix().get(i, j) ;
changes = temp * temp ;
- changes = sqrt(changes) ;
+ changes = vcl_sqrt(changes) ;
if ( changes > this->GetMinimalParametersChange() )
{
changed = true ;
diff --git a/Utilities/ITK/Code/Numerics/Statistics/itkGoodnessOfFitFunctionBase.txx b/Utilities/ITK/Code/Numerics/Statistics/itkGoodnessOfFitFunctionBase.txx
index 6c5c7de848..6ec31833f5 100644
--- a/Utilities/ITK/Code/Numerics/Statistics/itkGoodnessOfFitFunctionBase.txx
+++ b/Utilities/ITK/Code/Numerics/Statistics/itkGoodnessOfFitFunctionBase.txx
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkGoodnessOfFitFunctionBase.txx,v $
Language: C++
- Date: $Date: 2003/09/10 14:29:45 $
- Version: $Revision: 1.6 $
+ Date: $Date: 2006/03/19 04:37:20 $
+ Version: $Revision: 1.7 $
Copyright (c) Insight Software Consortium. All rights reserved.
See ITKCopyright.txt or http://www.itk.org/HTML/Copyright.htm for details.
@@ -27,7 +27,7 @@ GoodnessOfFitFunctionBase< TInputHistogram >
::GoodnessOfFitFunctionBase()
{
m_Epsilon = 1e-20 ;
- m_LogEpsilon = log(m_Epsilon) ;
+ m_LogEpsilon = vcl_log(m_Epsilon) ;
m_UseExpectedHistogram = true ;
m_ObservedHistogram = 0 ;
diff --git a/Utilities/ITK/Code/Numerics/Statistics/itkHistogram.txx b/Utilities/ITK/Code/Numerics/Statistics/itkHistogram.txx
index 91216bdd21..75881b3e62 100644
--- a/Utilities/ITK/Code/Numerics/Statistics/itkHistogram.txx
+++ b/Utilities/ITK/Code/Numerics/Statistics/itkHistogram.txx
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkHistogram.txx,v $
Language: C++
- Date: $Date: 2005/09/30 14:14:18 $
- Version: $Revision: 1.41 $
+ Date: $Date: 2006/06/09 19:11:56 $
+ Version: $Revision: 1.41.8.1 $
Copyright (c) Insight Software Consortium. All rights reserved.
See ITKCopyright.txt or http://www.itk.org/HTML/Copyright.htm for details.
@@ -137,85 +137,8 @@ inline const typename Histogram<TMeasurement, VMeasurementVectorSize,
Histogram<TMeasurement, VMeasurementVectorSize, TFrequencyContainer>
::GetIndex(const MeasurementVectorType& measurement) const
{
- // now using something similar to binary search to find
- // index.
- unsigned int dim ;
-
- int begin, mid, end ;
- MeasurementType median ;
- MeasurementType tempMeasurement ;
-
- for (dim = 0 ; dim < MeasurementVectorSize ; dim++)
- {
- tempMeasurement = measurement[dim] ;
- begin = 0 ;
- if (tempMeasurement < m_Min[dim][begin])
- {
- // one of measurement is below the minimum
- if(!m_ClipBinsAtEnds)
- {
- m_TempIndex[dim] = (long) 0 ;
- }
- else
- {
- // If clipping and this is below, then set to an illegal value
- // then set to m_Size[dim] to flag this as an illegal index;
- // Must be set to m_Size[dim] since -1 might not be a valid value.
- // This is why you must call IsIndexOutOfBounds before using the
- // index returned by this function.
- m_TempIndex[dim] = (long) m_Size[dim] ;
- itkWarningMacro(<<"One of the measurements is below the minimum");
- }
- }
-
- end = m_Min[dim].size() - 1 ;
- if (tempMeasurement >= m_Max[dim][end])
- {
- // one of measurement is above the maximum
- if(!m_ClipBinsAtEnds)
- {
- m_TempIndex[dim] = (long) m_Size[dim] - 1;
- }
- else
- {
- // If clipping and this is below, then set to an illegal value
- // then set to m_Size[dim] to flag this as an illegal index;
- // Must be set to m_Size[dim] since -1 might not be a valid value.
- // This is why you must call IsIndexOutOfBounds before using the
- // index returned by this function.
- m_TempIndex[dim] = (long) m_Size[dim] ;
- itkWarningMacro(<<"One of the measurements is above the maximum");
- }
- }
-
- mid = (end + 1) / 2 ;
- median = m_Min[dim][mid] ;
- while(true)
- {
- if (tempMeasurement < median )
- {
- end = mid - 1 ;
- }
- else if (tempMeasurement > median)
- {
- if (tempMeasurement < m_Max[dim][mid])
- {
- m_TempIndex[dim] = mid ;
- break ;
- }
-
- begin = mid + 1 ;
- }
- else
- {
- // measurement[dim] = m_Min[dim][med]
- m_TempIndex[dim] = mid ;
- break ;
- }
- mid = begin + (end - begin) / 2 ;
- median = m_Min[dim][mid] ;
- } // end of while
- } // end of for()
+ // Have this deprecated method call the un-deprecated one..
+ this->GetIndex( measurement, m_TempIndex );
return m_TempIndex;
}
@@ -242,16 +165,34 @@ bool Histogram<TMeasurement, VMeasurementVectorSize, TFrequencyContainer>
if (tempMeasurement < m_Min[dim][begin])
{
// one of measurement is below the minimum
- index[dim] = (long) m_Size[dim] ;
- return false;
+ // its ok if we extend the bins to infinity.. not ok if we don't
+ if(!m_ClipBinsAtEnds)
+ {
+ index[dim] = (long) 0 ;
+ continue;
+ }
+ else
+ { // set an illegal value and return 0
+ index[dim] = (long) m_Size[dim];
+ return false;
+ }
}
end = m_Min[dim].size() - 1 ;
if (tempMeasurement >= m_Max[dim][end])
{
- // one of measurement is above the maximum
- index[dim] = (long) m_Size[dim] ;
- return false;
+ // one of measurement is below the minimum
+ // its ok if we extend the bins to infinity.. not ok if we don't
+ if(!m_ClipBinsAtEnds)
+ {
+ index[dim] = (long) m_Size[dim]-1;
+ continue;
+ }
+ else
+ { // set an illegal value and return 0
+ index[dim] = (long) m_Size[dim];
+ return false;
+ }
}
mid = (end + 1) / 2 ;
diff --git a/Utilities/ITK/Code/Numerics/Statistics/itkHistogramToEntropyImageFilter.h b/Utilities/ITK/Code/Numerics/Statistics/itkHistogramToEntropyImageFilter.h
index ff7f153739..f2b47910a4 100644
--- a/Utilities/ITK/Code/Numerics/Statistics/itkHistogramToEntropyImageFilter.h
+++ b/Utilities/ITK/Code/Numerics/Statistics/itkHistogramToEntropyImageFilter.h
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkHistogramToEntropyImageFilter.h,v $
Language: C++
- Date: $Date: 2005/04/08 20:45:48 $
- Version: $Revision: 1.4 $
+ Date: $Date: 2006/03/19 04:37:20 $
+ 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.
@@ -72,13 +72,13 @@ public:
{
const double p = static_cast<OutputPixelType>(A) /
static_cast<OutputPixelType>(m_TotalFrequency);
- return static_cast<OutputPixelType>( (-1) * p * log(p) / log(2.0));
+ return static_cast<OutputPixelType>( (-1) * p * vcl_log(p) / vcl_log(2.0));
}
else
{
const double p = static_cast<OutputPixelType>(A+1) /
static_cast<OutputPixelType>(m_TotalFrequency);
- return static_cast<OutputPixelType>( (-1) * p * log(p) / log(2.0));
+ return static_cast<OutputPixelType>( (-1) * p * vcl_log(p) / vcl_log(2.0));
}
}
diff --git a/Utilities/ITK/Code/Numerics/Statistics/itkHistogramToLogProbabilityImageFilter.h b/Utilities/ITK/Code/Numerics/Statistics/itkHistogramToLogProbabilityImageFilter.h
index 61df35d997..e67cf5c063 100644
--- a/Utilities/ITK/Code/Numerics/Statistics/itkHistogramToLogProbabilityImageFilter.h
+++ b/Utilities/ITK/Code/Numerics/Statistics/itkHistogramToLogProbabilityImageFilter.h
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkHistogramToLogProbabilityImageFilter.h,v $
Language: C++
- Date: $Date: 2005/04/08 21:32:05 $
- Version: $Revision: 1.4 $
+ Date: $Date: 2006/03/19 04:37:20 $
+ 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.
@@ -66,12 +66,12 @@ public:
if( A )
{
return static_cast<OutputPixelType>(log( static_cast<OutputPixelType>(A) /
- static_cast<OutputPixelType>(m_TotalFrequency)) / log(2.0) );
+ static_cast<OutputPixelType>(m_TotalFrequency)) / vcl_log(2.0) );
}
else
{ // Check for Log 0. Always assume that the frequency is atleast 1.
return static_cast<OutputPixelType>(log( static_cast<OutputPixelType>(A+1) /
- static_cast<OutputPixelType>(m_TotalFrequency)) / log(2.0) );
+ static_cast<OutputPixelType>(m_TotalFrequency)) / vcl_log(2.0) );
}
}
diff --git a/Utilities/ITK/Code/Numerics/Statistics/itkListSampleToHistogramGenerator.h b/Utilities/ITK/Code/Numerics/Statistics/itkListSampleToHistogramGenerator.h
index c7c2ededf3..434159b8dc 100644
--- a/Utilities/ITK/Code/Numerics/Statistics/itkListSampleToHistogramGenerator.h
+++ b/Utilities/ITK/Code/Numerics/Statistics/itkListSampleToHistogramGenerator.h
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkListSampleToHistogramGenerator.h,v $
Language: C++
- Date: $Date: 2005/08/24 15:16:48 $
- Version: $Revision: 1.13 $
+ Date: $Date: 2006/03/14 22:52:18 $
+ Version: $Revision: 1.14 $
Copyright (c) Insight Software Consortium. All rights reserved.
See ITKCopyright.txt or http://www.itk.org/HTML/Copyright.htm for details.
@@ -22,6 +22,7 @@
#include "itkHistogram.h"
#include "itkStatisticsAlgorithm.h"
#include "itkDenseFrequencyContainer.h"
+#include "itkNumericTraits.h"
namespace itk{
namespace Statistics{
@@ -87,7 +88,11 @@ public:
itkStaticConstMacro(MeasurementVectorSize, unsigned int,
TMeasurementVectorLength);
- typedef Histogram< THistogramMeasurement,
+ /** Type needed for defining the limits of the histogram bins */
+ typedef typename NumericTraits< THistogramMeasurement >::RealType
+ HistogramMeasurementRealType;
+
+ typedef Histogram< HistogramMeasurementRealType,
itkGetStaticConstMacro(MeasurementVectorSize),
TFrequencyContainer > HistogramType ;
diff --git a/Utilities/ITK/Code/Numerics/Statistics/itkLogLikelihoodGoodnessOfFitFunction.txx b/Utilities/ITK/Code/Numerics/Statistics/itkLogLikelihoodGoodnessOfFitFunction.txx
index aa8be17598..00cac5cfbe 100644
--- a/Utilities/ITK/Code/Numerics/Statistics/itkLogLikelihoodGoodnessOfFitFunction.txx
+++ b/Utilities/ITK/Code/Numerics/Statistics/itkLogLikelihoodGoodnessOfFitFunction.txx
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkLogLikelihoodGoodnessOfFitFunction.txx,v $
Language: C++
- Date: $Date: 2004/09/11 00:15:16 $
- Version: $Revision: 1.10 $
+ Date: $Date: 2006/03/19 04:37:20 $
+ Version: $Revision: 1.11 $
Copyright (c) Insight Software Consortium. All rights reserved.
See ITKCopyright.txt or http://www.itk.org/HTML/Copyright.htm for details.
@@ -52,7 +52,7 @@ LogLikelihoodGoodnessOfFitFunction< TInputHistogram >
if ( ratio > this->GetEpsilon() && px > 0 )
{
- sum += px * log(ratio);
+ sum += px * vcl_log(ratio);
}
else
{
diff --git a/Utilities/ITK/Code/Numerics/Statistics/itkNormalVariateGenerator.cxx b/Utilities/ITK/Code/Numerics/Statistics/itkNormalVariateGenerator.cxx
index 7967af5d7c..4d9a3fbd87 100644
--- a/Utilities/ITK/Code/Numerics/Statistics/itkNormalVariateGenerator.cxx
+++ b/Utilities/ITK/Code/Numerics/Statistics/itkNormalVariateGenerator.cxx
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkNormalVariateGenerator.cxx,v $
Language: C++
- Date: $Date: 2004/04/19 19:33:48 $
- Version: $Revision: 1.14 $
+ Date: $Date: 2006/03/19 04:37:20 $
+ Version: $Revision: 1.15 $
Copyright (c) Insight Software Consortium. All rights reserved.
See ITKCopyright.txt or http://www.itk.org/HTML/Copyright.htm for details.
@@ -87,11 +87,11 @@ void NormalVariateGenerator::Initialize(int randomSeed)
// Z = (sqrt (1/2TLEN)) * A * (B + n)
// where:
// B = C / A.
-// We set chic1 = A * sqrt (0.5 / TLEN), chic2 = B
+// We set chic1 = A * vcl_sqrt(0.5 / TLEN), chic2 = B
fake = 1.0 + 0.125 / TLEN; // This is A
- chic2 = sqrt (2.0 * TLEN - fake*fake) / fake;
- chic1 = fake * sqrt (0.5 / TLEN);
+ chic2 = vcl_sqrt(2.0 * TLEN - fake*fake) / fake;
+ chic1 = fake * vcl_sqrt(0.5 / TLEN);
actualRSD = 0.0 ;
return;
@@ -324,16 +324,16 @@ double NormalVariateGenerator::FastNorm(void)
irs = (irs <= 0) ? ((irs << 1) ^ 333556017):(irs << 1);
r = irs + lseed;
if (r < 0) r = ~r;
- tz = -2.0 * log ((r + 0.5) * Rcons); /* Sum of squares */
+ tz = -2.0 * vcl_log((r + 0.5) * Rcons); /* Sum of squares */
ts += tz;
- tz = sqrt ( tz / tr );
+ tz = vcl_sqrt(tz / tr );
vec1 [p++] = (int) (Scale * tx * tz) ;
vec1 [p++] = (int) (Scale * ty * tz) ;
if (p < TLEN) goto nextpair;
/* Horrid, but good enough */
/* Calc correction factor to make sum of squares = TLEN */
ts = TLEN / ts; /* Should be close to 1.0 */
- tr = sqrt (ts);
+ tr = vcl_sqrt(ts);
for (p = 0; p < TLEN; p++) {
tx = vec1 [p] * tr;
vec1 [p] = (int) ((tx < 0.0) ? (tx - 0.5) : (tx + 0.5)) ;
@@ -347,7 +347,7 @@ double NormalVariateGenerator::FastNorm(void)
ts += (tx * tx);
}
/* Now ts should be Scale*Scale*TLEN or thereabouts */
- ts = sqrt (ts / (Scale * Scale * TLEN));
+ ts = vcl_sqrt(ts / (Scale * Scale * TLEN));
actualRSD = 1.0 / ts; /* Reciprocal of actual Standard Devtn */
goto startpass;
diff --git a/Utilities/ITK/Code/Numerics/Statistics/itkSampleToHistogramProjectionFilter.txx b/Utilities/ITK/Code/Numerics/Statistics/itkSampleToHistogramProjectionFilter.txx
index 976ecce612..8c6aa77d0d 100644
--- a/Utilities/ITK/Code/Numerics/Statistics/itkSampleToHistogramProjectionFilter.txx
+++ b/Utilities/ITK/Code/Numerics/Statistics/itkSampleToHistogramProjectionFilter.txx
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkSampleToHistogramProjectionFilter.txx,v $
Language: C++
-Date: $Date: 2005/07/26 15:55:05 $
-Version: $Revision: 1.13 $
+Date: $Date: 2006/03/19 04:37:20 $
+Version: $Revision: 1.14 $
Copyright (c) Insight Software Consortium. All rights reserved.
See ITKCopyright.txt or http://www.itk.org/HTML/Copyright.htm for details.
@@ -242,7 +242,7 @@ SampleToHistogramProjectionFilter< TInputSample, THistogramMeasurement >
{
if (minWeight < 1.0)
{
- minWeight = 1.0 / (1.0 + exp(-minWeight / scale)) ;
+ minWeight = 1.0 / (1.0 + vcl_exp(-minWeight / scale)) ;
}
else
{
@@ -251,7 +251,7 @@ SampleToHistogramProjectionFilter< TInputSample, THistogramMeasurement >
if (maxWeight < 1.0)
{
- maxWeight = 1.0 / (1.0 + exp(-maxWeight / scale)) ;
+ maxWeight = 1.0 / (1.0 + vcl_exp(-maxWeight / scale)) ;
}
else
{
@@ -295,7 +295,7 @@ SampleToHistogramProjectionFilter< TInputSample, THistogramMeasurement >
if (m_HistogramBinOverlap)
{
- scale = log(1.0 + m_HistogramBinOverlap / 10.0 ) ;
+ scale = vcl_log(1.0 + m_HistogramBinOverlap / 10.0 ) ;
}
typename HistogramType::InstanceIdentifier binId ;
@@ -332,7 +332,7 @@ SampleToHistogramProjectionFilter< TInputSample, THistogramMeasurement >
}
marginalDistance =
- sqrt(vnl_math_abs(squaredDistance - dotProduct * dotProduct)) /
+ vcl_sqrt(vnl_math_abs(squaredDistance - dotProduct * dotProduct)) /
((*m_StandardDeviation) * extent) ;
dotProduct /= (*m_StandardDeviation) ;
@@ -360,7 +360,7 @@ SampleToHistogramProjectionFilter< TInputSample, THistogramMeasurement >
if ( marginalDistance > -1.0 )
{
marginalDistance =
- 1.0 / (1.0 + exp(-marginalDistance / scale)) ;
+ 1.0 / (1.0 + vcl_exp(-marginalDistance / scale)) ;
for (binId = 0 ; binId <= (numberOfBins / 2UL) ;
binId++)
diff --git a/Utilities/ITK/Code/Numerics/Statistics/itkScalarImageToHistogramGenerator.h b/Utilities/ITK/Code/Numerics/Statistics/itkScalarImageToHistogramGenerator.h
index ef40a52b89..65076814c6 100644
--- a/Utilities/ITK/Code/Numerics/Statistics/itkScalarImageToHistogramGenerator.h
+++ b/Utilities/ITK/Code/Numerics/Statistics/itkScalarImageToHistogramGenerator.h
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkScalarImageToHistogramGenerator.h,v $
Language: C++
- Date: $Date: 2005/08/24 15:16:56 $
- Version: $Revision: 1.7 $
+ Date: $Date: 2006/03/14 22:01:52 $
+ Version: $Revision: 1.8 $
Copyright (c) Insight Software Consortium. All rights reserved.
See ITKCopyright.txt or http://www.itk.org/HTML/Copyright.htm for details.
@@ -48,10 +48,11 @@ public:
> AdaptorType;
typedef typename AdaptorType::Pointer AdaptorPointer;
typedef typename ImageType::PixelType PixelType;
+ typedef typename NumericTraits< PixelType >::RealType RealPixelType;
typedef itk::Statistics::ListSampleToHistogramGenerator<
AdaptorType,
- PixelType,
+ RealPixelType,
DenseFrequencyContainer
> GeneratorType;
@@ -80,6 +81,11 @@ public:
/** Set marginal scale value to be passed to the histogram generator */
void SetMarginalScale( double marginalScale );
+ /** Set the minimum value from which the bins will be computed */
+ void SetHistogramMin( RealPixelType minimumValue );
+
+ /** Set the maximum value from which the bins will be computed */
+ void SetHistogramMax( RealPixelType maximumValue );
protected:
ScalarImageToHistogramGenerator();
diff --git a/Utilities/ITK/Code/Numerics/Statistics/itkScalarImageToHistogramGenerator.txx b/Utilities/ITK/Code/Numerics/Statistics/itkScalarImageToHistogramGenerator.txx
index 21986bc2d2..8a1109360e 100644
--- a/Utilities/ITK/Code/Numerics/Statistics/itkScalarImageToHistogramGenerator.txx
+++ b/Utilities/ITK/Code/Numerics/Statistics/itkScalarImageToHistogramGenerator.txx
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkScalarImageToHistogramGenerator.txx,v $
Language: C++
- Date: $Date: 2003/12/05 18:47:38 $
- Version: $Revision: 1.1 $
+ Date: $Date: 2006/03/14 22:01:52 $
+ Version: $Revision: 1.2 $
Copyright (c) Insight Software Consortium. All rights reserved.
See ITKCopyright.txt or http://www.itk.org/HTML/Copyright.htm for details.
@@ -75,6 +75,30 @@ ScalarImageToHistogramGenerator< TImage >
}
+template < class TImage >
+void
+ScalarImageToHistogramGenerator< TImage >
+::SetHistogramMin( RealPixelType minimumValue )
+{
+ typedef typename GeneratorType::MeasurementVectorType MeasurementVectorType;
+ MeasurementVectorType minVector;
+ minVector[0] = minimumValue;
+ m_HistogramGenerator->SetHistogramMin( minVector );
+}
+
+
+template < class TImage >
+void
+ScalarImageToHistogramGenerator< TImage >
+::SetHistogramMax( RealPixelType maximumValue )
+{
+ typedef typename GeneratorType::MeasurementVectorType MeasurementVectorType;
+ MeasurementVectorType maxVector;
+ maxVector[0] = maximumValue;
+ m_HistogramGenerator->SetHistogramMax( maxVector );
+}
+
+
template < class TImage >
void
diff --git a/Utilities/ITK/Code/Numerics/itkCumulativeGaussianCostFunction.cxx b/Utilities/ITK/Code/Numerics/itkCumulativeGaussianCostFunction.cxx
index 2ac0bb7458..8485f9380c 100644
--- a/Utilities/ITK/Code/Numerics/itkCumulativeGaussianCostFunction.cxx
+++ b/Utilities/ITK/Code/Numerics/itkCumulativeGaussianCostFunction.cxx
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkCumulativeGaussianCostFunction.cxx,v $
Language: C++
- Date: $Date: 2003/12/15 14:13:19 $
- Version: $Revision: 1.11 $
+ Date: $Date: 2006/03/19 04:37:19 $
+ Version: $Revision: 1.12 $
Copyright (c) Insight Software Consortium. All rights reserved.
See ITKCopyright.txt or http://www.itk.org/HTML/Copyright.htm for details.
@@ -66,7 +66,7 @@ CumulativeGaussianCostFunction
double fitError = 0.0;
for(int i = 0; i < (int)(numberOfElements); i++)
{
- fitError += pow((setTestArray->get(i) - m_OriginalDataArray->get(i)), 2);
+ fitError += vcl_pow((setTestArray->get(i) - m_OriginalDataArray->get(i)), 2);
}
return(sqrt((1/numberOfElements) * fitError));
}
@@ -148,7 +148,7 @@ CumulativeGaussianCostFunction
::GetValue( const ParametersType & parameters ) const
{
for(unsigned int i = 0; i < m_RangeDimension; i++)
- m_Measure[i] = parameters.get(2) + ((parameters.get(3)-parameters.get(2))*(EvaluateCumulativeGaussian((i-parameters.get(0))/(parameters.get(1)*sqrt(2.0)))+1)/2);
+ m_Measure[i] = parameters.get(2) + ((parameters.get(3)-parameters.get(2))*(EvaluateCumulativeGaussian((i-parameters.get(0))/(parameters.get(1)*vcl_sqrt(2.0)))+1)/2);
return m_Measure;
}
@@ -160,7 +160,7 @@ CumulativeGaussianCostFunction
m_MeasurePointer->SetSize(m_RangeDimension);
for(unsigned int i = 0; i < m_RangeDimension; i++)
- m_MeasurePointer->put(i, parameters.get(2) + ((parameters.get(3)-parameters.get(2))*(EvaluateCumulativeGaussian((i-parameters.get(0))/(parameters.get(1)*sqrt(2.0)))+1)/2));
+ m_MeasurePointer->put(i, parameters.get(2) + ((parameters.get(3)-parameters.get(2))*(EvaluateCumulativeGaussian((i-parameters.get(0))/(parameters.get(1)*vcl_sqrt(2.0)))+1)/2));
return m_MeasurePointer;
}
diff --git a/Utilities/ITK/Code/Numerics/itkCumulativeGaussianOptimizer.cxx b/Utilities/ITK/Code/Numerics/itkCumulativeGaussianOptimizer.cxx
index c0a04c6a15..8849ac7f7f 100644
--- a/Utilities/ITK/Code/Numerics/itkCumulativeGaussianOptimizer.cxx
+++ b/Utilities/ITK/Code/Numerics/itkCumulativeGaussianOptimizer.cxx
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkCumulativeGaussianOptimizer.cxx,v $
Language: C++
- Date: $Date: 2003/12/15 14:13:19 $
- Version: $Revision: 1.14 $
+ Date: $Date: 2006/03/19 04:37:20 $
+ Version: $Revision: 1.15 $
Copyright (c) Insight Software Consortium. All rights reserved.
See ITKCopyright.txt or http://www.itk.org/HTML/Copyright.htm for details.
@@ -59,7 +59,7 @@ CumulativeGaussianOptimizer
for(int i=0; i<(int)(extendedArray->GetNumberOfElements()); i++)
{
- extendedArray->put(i, amplitude * exp( - ( pow((i-mean),2) / (2*pow(sd,2)) ) ));
+ extendedArray->put(i, amplitude * vcl_exp(- ( vcl_pow((i-mean),2) / (2*vcl_pow(sd,2)) ) ));
}
// Then insert the originalArray over the middle section of extendedArray.
for(int i=0; i<(int)(originalArray->GetNumberOfElements()); i++)
@@ -162,15 +162,15 @@ void CumulativeGaussianOptimizer
// Calculate the standard deviation
for(int i = 0; i < (int)(array->GetNumberOfElements()); i++)
- m_ComputedStandardDeviation += array->get(i) * pow( (i - m_ComputedMean), 2);
+ m_ComputedStandardDeviation += array->get(i) * vcl_pow((i - m_ComputedMean), 2);
- m_ComputedStandardDeviation = sqrt( m_ComputedStandardDeviation/sum );
+ m_ComputedStandardDeviation = vcl_sqrt(m_ComputedStandardDeviation/sum );
// For the ERF, sum is the difference between the lower and upper intensities.
m_ComputedTransitionHeight = sum;
// Calculate the amplitude.
- m_ComputedAmplitude = sum / (m_ComputedStandardDeviation * sqrt(2*3.14159265));
+ m_ComputedAmplitude = sum / (m_ComputedStandardDeviation * vcl_sqrt(2*3.14159265));
}
void
@@ -208,7 +208,7 @@ CumulativeGaussianOptimizer
if( i < startingPointForInsertion ||
i >= startingPointForInsertion + (int)(originalArray->GetNumberOfElements()) )
{
- extendedArray->put(i, amplitude * exp(-(pow((i - mean),2) / (2 * pow(sd,2)))));
+ extendedArray->put(i, amplitude * vcl_exp(-(pow((i - mean),2) / (2 * vcl_pow(sd,2)))));
}
}
return extendedArray;
@@ -254,7 +254,7 @@ CumulativeGaussianOptimizer
// Generate new Gaussian array with final parameters.
for(int i = 0; i < sampledGaussianArraySize; i++)
- sampledGaussianArray->put(i, m_ComputedAmplitude * exp( - ( pow((i-m_ComputedMean),2) / (2*pow(m_ComputedStandardDeviation,2)) ) ));
+ sampledGaussianArray->put(i, m_ComputedAmplitude * vcl_exp(- ( vcl_pow((i-m_ComputedMean),2) / (2*vcl_pow(m_ComputedStandardDeviation,2)) ) ));
// Add 0.5 to the mean of the sampled Gaussian curve to make up for the 0.5
// shift during derivation, then take the integral of the Gaussian sample
diff --git a/Utilities/ITK/Code/Numerics/itkFRPROptimizer.cxx b/Utilities/ITK/Code/Numerics/itkFRPROptimizer.cxx
index b1802f7afa..d1e36b838d 100644
--- a/Utilities/ITK/Code/Numerics/itkFRPROptimizer.cxx
+++ b/Utilities/ITK/Code/Numerics/itkFRPROptimizer.cxx
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkFRPROptimizer.cxx,v $
Language: C++
- Date: $Date: 2005/04/27 15:08:57 $
- Version: $Revision: 1.5 $
+ Date: $Date: 2006/05/25 13:41:42 $
+ Version: $Revision: 1.7 $
Copyright (c) Insight Software Consortium. All rights reserved.
See ITKCopyright.txt or http://www.itk.org/HTML/Copyright.htm for details.
@@ -19,14 +19,16 @@
#include "itkFRPROptimizer.h"
-#define FRPR_TINY 1e-20
namespace itk
{
+const double FRPR_TINY = 1e-20;
+
FRPROptimizer
::FRPROptimizer()
{
+ m_OptimizationType = PolakRibiere;
}
FRPROptimizer
@@ -94,18 +96,20 @@ FRPROptimizer
this->SetSpaceDimension(m_CostFunction->GetNumberOfParameters());
+ const unsigned int SpaceDimension = this->GetSpaceDimension();
+
double gg, gam, dgg;
- FRPROptimizer::ParametersType g(this->GetSpaceDimension());
- FRPROptimizer::ParametersType h(this->GetSpaceDimension());
- FRPROptimizer::ParametersType xi(this->GetSpaceDimension());
+ FRPROptimizer::ParametersType g( SpaceDimension );
+ FRPROptimizer::ParametersType h( SpaceDimension );
+ FRPROptimizer::ParametersType xi( SpaceDimension );
- FRPROptimizer::ParametersType p(this->GetSpaceDimension());
+ FRPROptimizer::ParametersType p( SpaceDimension );
p = this->GetInitialPosition();
double fp;
this->GetValueAndDerivative(p, &fp, &xi);
- for(i=0; i<this->GetSpaceDimension(); i++)
+ for( i = 0; i < SpaceDimension ; i++ )
{
g[i] = -xi[i];
xi[i] = g[i];
@@ -127,7 +131,7 @@ FRPROptimizer
if ( 2.0 * vcl_abs(fret - fp) <=
this->GetValueTolerance() * (vcl_abs(fret)+ vcl_abs(fp) + FRPR_TINY) )
{
- if(limitCount < this->GetSpaceDimension())
+ if( limitCount < SpaceDimension )
{
this->GetValueAndDerivative(p, &fp, &xi);
xi[limitCount] = 1;
@@ -148,11 +152,22 @@ FRPROptimizer
gg = 0.0;
dgg = 0.0;
- for(i=0; i<this->GetSpaceDimension(); i++)
+
+ if( m_OptimizationType == PolakRibiere )
{
- gg += g[i] * g[i];
- //dgg += xi[i] * xi[i]; // Uncomment for Fletcher-Reeves
- dgg += (xi[i] + g[i]) * xi[i]; // Uncoment for Polak-Ribiere
+ for( i=0; i< SpaceDimension ; i++ )
+ {
+ gg += g[i] * g[i];
+ dgg += (xi[i] + g[i]) * xi[i];
+ }
+ }
+ if( m_OptimizationType == FletchReeves )
+ {
+ for( i=0; i< SpaceDimension ; i++ )
+ {
+ gg += g[i] * g[i];
+ dgg += xi[i] * xi[i];
+ }
}
if(gg == 0)
@@ -163,7 +178,7 @@ FRPROptimizer
}
gam = dgg/gg;
- for(i=0; i<this->GetSpaceDimension(); i++)
+ for( i = 0; i < SpaceDimension ; i++)
{
g[i] = -xi[i];
xi[i] = g[i] + gam * h[i];
@@ -178,6 +193,28 @@ FRPROptimizer
this->InvokeEvent( EndEvent() );
}
+/**
+ *
+ */
+void
+FRPROptimizer
+::SetToPolakRibiere()
+{
+ m_OptimizationType = PolakRibiere;
+}
+
+
+/**
+ *
+ */
+void
+FRPROptimizer
+::SetToFletchReeves()
+{
+ m_OptimizationType = FletchReeves;
+}
+
+
/**
*
*/
@@ -186,6 +223,8 @@ FRPROptimizer
::PrintSelf(std::ostream& os, Indent indent) const
{
Superclass::PrintSelf(os,indent);
+ os << indent << "Optimization Type = " << m_OptimizationType << std::endl;
+ os << indent << "0=FletchReeves, 1=PolakRibiere" << std::endl;
}
} // end of namespace itk
diff --git a/Utilities/ITK/Code/Numerics/itkFRPROptimizer.h b/Utilities/ITK/Code/Numerics/itkFRPROptimizer.h
index 24fd9e1953..54adc0c7c3 100644
--- a/Utilities/ITK/Code/Numerics/itkFRPROptimizer.h
+++ b/Utilities/ITK/Code/Numerics/itkFRPROptimizer.h
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkFRPROptimizer.h,v $
Language: C++
- Date: $Date: 2005/03/22 00:13:48 $
- Version: $Revision: 1.1 $
+ Date: $Date: 2006/05/25 13:41:42 $
+ Version: $Revision: 1.2 $
Copyright (c) Insight Software Consortium. All rights reserved.
See ITKCopyright.txt or http://www.itk.org/HTML/Copyright.htm for details.
@@ -71,11 +71,17 @@ public:
typedef CostFunctionType::Pointer CostFunctionPointer;
/** Start optimization. */
- void StartOptimization() ;
+ void StartOptimization();
+ /** Set it to the Fletch-Reeves optimizer */
+ void SetToFletchReeves();
+
+ /** Set it to the Fletch-Reeves optimizer */
+ void SetToPolakRibiere();
+
protected:
FRPROptimizer() ;
- virtual ~FRPROptimizer() ;
+ virtual ~FRPROptimizer();
void PrintSelf(std::ostream& os, Indent indent) const;
@@ -92,6 +98,14 @@ protected:
private:
FRPROptimizer(const FRPROptimizer&) ; // not implemented
+ typedef enum
+ {
+ FletchReeves,
+ PolakRibiere
+ } OptimizationType;
+
+ OptimizationType m_OptimizationType;
+
} ; // end of class
} // end of namespace itk
diff --git a/Utilities/ITK/Code/Numerics/itkMultipleValuedNonLinearVnlOptimizer.cxx b/Utilities/ITK/Code/Numerics/itkMultipleValuedNonLinearVnlOptimizer.cxx
index ffc3e5744b..ea21ff5410 100644
--- a/Utilities/ITK/Code/Numerics/itkMultipleValuedNonLinearVnlOptimizer.cxx
+++ b/Utilities/ITK/Code/Numerics/itkMultipleValuedNonLinearVnlOptimizer.cxx
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkMultipleValuedNonLinearVnlOptimizer.cxx,v $
Language: C++
- Date: $Date: 2005/03/09 03:41:38 $
- Version: $Revision: 1.11 $
+ Date: $Date: 2006/05/12 12:58:27 $
+ Version: $Revision: 1.12 $
Copyright (c) Insight Software Consortium. All rights reserved.
See ITKCopyright.txt or http://www.itk.org/HTML/Copyright.htm for details.
@@ -30,6 +30,12 @@ MultipleValuedNonLinearVnlOptimizer
{
m_CostFunctionAdaptor = 0;
m_UseGradient = true;
+ m_Command = CommandType::New();
+ m_Command->SetCallbackFunction( this,
+ &MultipleValuedNonLinearVnlOptimizer::IterationReport );
+ m_CachedValue.Fill(0);
+ m_CachedCurrentPosition.Fill(0);
+ m_CachedDerivative.Fill(0);
}
@@ -68,6 +74,7 @@ MultipleValuedNonLinearVnlOptimizer
this->SetUseCostFunctionGradient(m_UseGradient);
+ m_CostFunctionAdaptor->AddObserver( IterationEvent(), m_Command );
}
@@ -131,6 +138,42 @@ MultipleValuedNonLinearVnlOptimizer
+/** The purpose of this method is to get around the lack of iteration reporting
+ * in VNL optimizers. By interfacing directly with the ITK cost function
+ * adaptor we are generating here Iteration Events. Note the iteration events
+ * here are produce PER EVALUATION of the metric, not per real iteration of the
+ * vnl optimizer. Optimizers that evaluate the metric multiple times at each
+ * iteration will generate a lot more of Iteration events here. */
+void
+MultipleValuedNonLinearVnlOptimizer
+::IterationReport( const EventObject & event )
+{
+ const CostFunctionAdaptorType * adaptor = this->GetCostFunctionAdaptor();
+ m_CachedValue = adaptor->GetCachedValue();
+ m_CachedDerivative = adaptor->GetCachedDerivative();
+ m_CachedCurrentPosition = adaptor->GetCachedCurrentParameters();
+ this->InvokeEvent( event );
+}
+
+
+/**
+ * PrintSelf
+ */
+void
+MultipleValuedNonLinearVnlOptimizer
+::PrintSelf(std::ostream& os, Indent indent) const
+{
+ Superclass::PrintSelf( os, indent );
+ os << indent << "Cached Value: " << m_CachedValue << std::endl;
+ os << indent << "Cached Derivative: " << m_CachedDerivative << std::endl;
+ os << indent << "Cached current positiion: "
+ << m_CachedCurrentPosition << std::endl;
+ os << "Command observer " << m_Command.GetPointer() << std::endl;
+ os << "Cost Function adaptor" << m_CostFunctionAdaptor << std::endl;
+}
+
+
+
} // end namespace itk
diff --git a/Utilities/ITK/Code/Numerics/itkMultipleValuedNonLinearVnlOptimizer.h b/Utilities/ITK/Code/Numerics/itkMultipleValuedNonLinearVnlOptimizer.h
index ab137c9b85..d365e813f5 100644
--- a/Utilities/ITK/Code/Numerics/itkMultipleValuedNonLinearVnlOptimizer.h
+++ b/Utilities/ITK/Code/Numerics/itkMultipleValuedNonLinearVnlOptimizer.h
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkMultipleValuedNonLinearVnlOptimizer.h,v $
Language: C++
- Date: $Date: 2005/09/24 21:20:58 $
- Version: $Revision: 1.23 $
+ Date: $Date: 2006/05/12 12:58:27 $
+ Version: $Revision: 1.24 $
Copyright (c) Insight Software Consortium. All rights reserved.
See ITKCopyright.txt or http://www.itk.org/HTML/Copyright.htm for details.
@@ -19,6 +19,7 @@
#include "itkMultipleValuedNonLinearOptimizer.h"
#include "itkMultipleValuedVnlCostFunctionAdaptor.h"
+#include "itkCommand.h"
namespace itk
@@ -67,10 +68,21 @@ public:
void UseCostFunctionGradientOff() { this->SetUseCostFunctionGradient( false ); };
bool GetUseCostFunctionGradient() const;
+ /** Return Cached Values. These method have the advantage of not triggering a
+ * recomputation of the metric value, but it has the disadvantage of returning
+ * a value that may not be the one corresponding to the current parameters. For
+ * GUI update purposes, this method is a good option, for mathematical
+ * validation you should rather call GetValue(). */
+ itkGetConstReferenceMacro(CachedValue, MeasureType);
+ itkGetConstReferenceMacro(CachedDerivative, DerivativeType);
+ itkGetConstReferenceMacro(CachedCurrentPosition, ParametersType);
+
protected:
MultipleValuedNonLinearVnlOptimizer();
virtual ~MultipleValuedNonLinearVnlOptimizer();
+ void PrintSelf(std::ostream& os, Indent indent) const;
+
typedef MultipleValuedVnlCostFunctionAdaptor CostFunctionAdaptorType;
void SetCostFunctionAdaptor( CostFunctionAdaptorType * adaptor );
@@ -81,12 +93,28 @@ protected:
* correctness in vnl cost_functions and optimizers */
CostFunctionAdaptorType * GetNonConstCostFunctionAdaptor( void ) const;
+ /** Command observer that will interact with the ITK-VNL cost-function
+ * adaptor in order to generate iteration events. This will allow to overcome
+ * the limitation of VNL optimizers not offering callbacks for every
+ * iteration */
+ typedef ReceptorMemberCommand< Self > CommandType;
+
private:
MultipleValuedNonLinearVnlOptimizer(const Self&); //purposely not implemented
void operator=(const Self&); //purposely not implemented
+ /** Callback function for the Command Observer */
+ void IterationReport( const EventObject & event );
+
CostFunctionAdaptorType * m_CostFunctionAdaptor;
bool m_UseGradient;
+
+ CommandType::Pointer m_Command;
+
+ mutable ParametersType m_CachedCurrentPosition;
+ mutable MeasureType m_CachedValue;
+ mutable DerivativeType m_CachedDerivative;
+
};
} // end namespace itk
diff --git a/Utilities/ITK/Code/Numerics/itkMultipleValuedVnlCostFunctionAdaptor.cxx b/Utilities/ITK/Code/Numerics/itkMultipleValuedVnlCostFunctionAdaptor.cxx
index 9d682bc152..61d4c48df8 100644
--- a/Utilities/ITK/Code/Numerics/itkMultipleValuedVnlCostFunctionAdaptor.cxx
+++ b/Utilities/ITK/Code/Numerics/itkMultipleValuedVnlCostFunctionAdaptor.cxx
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkMultipleValuedVnlCostFunctionAdaptor.cxx,v $
Language: C++
- Date: $Date: 2004/08/15 12:26:38 $
- Version: $Revision: 1.10 $
+ Date: $Date: 2006/05/12 12:58:27 $
+ Version: $Revision: 1.12 $
Copyright (c) Insight Software Consortium. All rights reserved.
See ITKCopyright.txt or http://www.itk.org/HTML/Copyright.htm for details.
@@ -29,6 +29,7 @@ MultipleValuedVnlCostFunctionAdaptor
vnl_least_squares_function(spaceDimension,numberOfValues)
{
m_ScalesInitialized =false;
+ m_Reporter = Object::New();
}
@@ -71,6 +72,11 @@ MultipleValuedVnlCostFunctionAdaptor
measures = m_CostFunction->GetValue( parameters );
+ // Notify observers. This is used for overcoming the limitaion of VNL
+ // optimizers of not providing callbacks per iteration.
+ m_CachedValue = measures;
+ m_CachedCurrentParameters = parameters;
+ this->ReportIteration( FunctionEvaluationIterationEvent() );
}
@@ -114,7 +120,7 @@ MultipleValuedVnlCostFunctionAdaptor
void
MultipleValuedVnlCostFunctionAdaptor
::compute( const InternalParametersType & x,
- InternalMeasureType * f,
+ InternalMeasureType * ff,
InternalDerivativeType * g )
{
// delegate the computation to the CostFunction
@@ -132,11 +138,19 @@ MultipleValuedVnlCostFunctionAdaptor
parameters.SetData(const_cast<double*>(x.data_block()));
}
- *f = static_cast<InternalMeasureType>(
+ *ff = static_cast<InternalMeasureType>(
m_CostFunction->GetValue( parameters ) );
m_CostFunction->GetDerivative( parameters, externalGradient );
ConvertExternalToInternalGradient( externalGradient, *g );
+
+ // Notify observers. This is used for overcoming the limitaion of VNL
+ // optimizers of not providing callbacks per iteration.
+ // Note that m_CachedDerivative is already loaded in the GetDerivative() above.
+ m_CachedValue = *ff;
+ m_CachedCurrentParameters = parameters;
+ this->ReportIteration( FunctionAndGradientEvaluationIterationEvent() );
+
}
/** Convert external derviative measures into internal type */
@@ -196,6 +210,53 @@ MultipleValuedVnlCostFunctionAdaptor
+/** This method reports iterations events. It is intended to
+ * help monitoring the progress of the optimization process. */
+void
+MultipleValuedVnlCostFunctionAdaptor
+::ReportIteration( const EventObject & event ) const
+{
+ this->m_Reporter->InvokeEvent( event );
+}
+
+
+
+/** Connects a Command/Observer to the internal reporter class.
+ * This is useful for reporting iteration event to potential observers. */
+unsigned long
+MultipleValuedVnlCostFunctionAdaptor
+::AddObserver(const EventObject & event, Command * command) const
+{
+ return m_Reporter->AddObserver( event, command );
+}
+
+
+/** Return the cached value of the cost function */
+const MultipleValuedVnlCostFunctionAdaptor::MeasureType &
+MultipleValuedVnlCostFunctionAdaptor
+::GetCachedValue() const
+{
+ return m_CachedValue;
+}
+
+
+/** Return the cached value of the cost function derivative */
+const MultipleValuedVnlCostFunctionAdaptor::DerivativeType &
+MultipleValuedVnlCostFunctionAdaptor
+::GetCachedDerivative() const
+{
+ return m_CachedDerivative;
+}
+
+/** Return the cached value of the parameters used for computing the function */
+const MultipleValuedVnlCostFunctionAdaptor::ParametersType &
+MultipleValuedVnlCostFunctionAdaptor
+::GetCachedCurrentParameters() const
+{
+ return m_CachedCurrentParameters;
+}
+
+
} // end namespace itk
diff --git a/Utilities/ITK/Code/Numerics/itkMultipleValuedVnlCostFunctionAdaptor.h b/Utilities/ITK/Code/Numerics/itkMultipleValuedVnlCostFunctionAdaptor.h
index 4556dfb4b5..2a6f2e4230 100644
--- a/Utilities/ITK/Code/Numerics/itkMultipleValuedVnlCostFunctionAdaptor.h
+++ b/Utilities/ITK/Code/Numerics/itkMultipleValuedVnlCostFunctionAdaptor.h
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkMultipleValuedVnlCostFunctionAdaptor.h,v $
Language: C++
- Date: $Date: 2004/03/14 17:43:02 $
- Version: $Revision: 1.5 $
+ Date: $Date: 2006/05/12 12:58:27 $
+ Version: $Revision: 1.6 $
Copyright (c) Insight Software Consortium. All rights reserved.
See ITKCopyright.txt or http://www.itk.org/HTML/Copyright.htm for details.
@@ -105,12 +105,38 @@ public:
/** Set current parameters scaling. */
void SetScales(const ScalesType & scales);
+ /** This AddObserver method allows to simulate that this class derives from
+ * an itkObject for the purpose of reporting iteration events. The goal of
+ * this method is to allow ITK-vnl optimizer adaptors to get iteration events
+ * despite the fact that VNL does not provide callbacks. */
+ unsigned long AddObserver(const EventObject & event, Command *) const;
+
+ /** Return the value of the last evaluation to the value of the cost function.
+ * Note that this method DOES NOT triggers a computation of the function or
+ * the derivatives, it only returns previous values. Therefore the values here
+ * are only valid after you invoke the f() or gradf() methods. */
+ const MeasureType & GetCachedValue() const;
+ const DerivativeType & GetCachedDerivative() const;
+ const ParametersType & GetCachedCurrentParameters() const;
+
+protected:
+
+ /** This method is intended to be called by the derived classes in order to
+ * notify of an iteration event to any Command/Observers */
+ void ReportIteration( const EventObject & event ) const;
+
+
private:
MultipleValuedCostFunction::Pointer m_CostFunction;
+
bool m_ScalesInitialized;
ScalesType m_Scales;
+ Object::Pointer m_Reporter;
+ mutable MeasureType m_CachedValue;
+ mutable DerivativeType m_CachedDerivative;
+ mutable ParametersType m_CachedCurrentParameters;
}; // end of Class CostFunction
diff --git a/Utilities/ITK/Code/Numerics/itkOnePlusOneEvolutionaryOptimizer.cxx b/Utilities/ITK/Code/Numerics/itkOnePlusOneEvolutionaryOptimizer.cxx
index d3aeb62715..ee3cc26025 100644
--- a/Utilities/ITK/Code/Numerics/itkOnePlusOneEvolutionaryOptimizer.cxx
+++ b/Utilities/ITK/Code/Numerics/itkOnePlusOneEvolutionaryOptimizer.cxx
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkOnePlusOneEvolutionaryOptimizer.cxx,v $
Language: C++
- Date: $Date: 2005/03/07 03:39:58 $
- Version: $Revision: 1.32 $
+ Date: $Date: 2006/03/19 04:37:20 $
+ Version: $Revision: 1.33 $
Copyright (c) Insight Software Consortium. All rights reserved.
See ITKCopyright.txt or http://www.itk.org/HTML/Copyright.htm for details.
@@ -32,7 +32,7 @@ OnePlusOneEvolutionaryOptimizer
m_Initialized = false ;
m_GrowthFactor = 1.05 ;
- m_ShrinkFactor = pow(m_GrowthFactor, -0.25) ;
+ m_ShrinkFactor = vcl_pow(m_GrowthFactor, -0.25) ;
m_InitialRadius = 1.01 ;
m_MaximumIteration = 100 ;
m_Stop = false ;
@@ -69,7 +69,7 @@ OnePlusOneEvolutionaryOptimizer
m_GrowthFactor = grow ;
if (shrink == -1)
- m_ShrinkFactor = pow(m_GrowthFactor, -0.25) ;
+ m_ShrinkFactor = vcl_pow(m_GrowthFactor, -0.25) ;
else
m_ShrinkFactor = shrink ;
}
diff --git a/Utilities/ITK/Code/Numerics/itkSPSAOptimizer.cxx b/Utilities/ITK/Code/Numerics/itkSPSAOptimizer.cxx
index 95a9e35301..bc69ac9ac7 100644
--- a/Utilities/ITK/Code/Numerics/itkSPSAOptimizer.cxx
+++ b/Utilities/ITK/Code/Numerics/itkSPSAOptimizer.cxx
@@ -6,7 +6,6 @@
#include "itkEventObject.h"
#include "itkExceptionObject.h"
-#include "vnl/vnl_sample.h"
#include "vnl/vnl_math.h"
@@ -37,6 +36,7 @@ namespace itk
m_A = m_MaximumNumberOfIterations / 10;
m_Alpha = 0.602;
m_Gamma = 0.101;
+ m_Generator = Statistics::MersenneTwisterRandomVariateGenerator::New();
} // end Constructor
@@ -336,7 +336,9 @@ namespace itk
for ( unsigned int j = 0; j < spaceDimension; j++ )
{
/** Generate randomly -1 or 1. */
- m_Delta[ j ] = 2 * vnl_math_rnd( vnl_sample_uniform(0.0f,1.0f) ) - 1;
+ // m_Delta[ j ] = 2 * vnl_math_rnd( vnl_sample_uniform(0.0f,1.0f) ) - 1;
+
+ m_Delta[ j ] = 2 * vnl_math_rnd ( this->m_Generator->GetUniformVariate (0.0f, 1.0f) ) - 1;
/**
* Take scales into account. The perturbation of a parameter that has a
@@ -491,7 +493,7 @@ namespace itk
this->ComputeGradient(initialPosition, m_Gradient);
for ( unsigned int j = 0; j < spaceDimension; j++ )
{
- averageAbsoluteGradient[j] += fabs(m_Gradient[j]);
+ averageAbsoluteGradient[j] += vcl_fabs(m_Gradient[j]);
}
} // end for ++n
averageAbsoluteGradient /= static_cast<double>(numberOfGradientEstimates);
diff --git a/Utilities/ITK/Code/Numerics/itkSPSAOptimizer.h b/Utilities/ITK/Code/Numerics/itkSPSAOptimizer.h
index 9ee17c1d48..4a78354198 100644
--- a/Utilities/ITK/Code/Numerics/itkSPSAOptimizer.h
+++ b/Utilities/ITK/Code/Numerics/itkSPSAOptimizer.h
@@ -2,6 +2,7 @@
#define __itkSPSAOptimizer_h
#include "itkSingleValuedNonLinearOptimizer.h"
+#include "itkMersenneTwisterRandomVariateGenerator.h"
namespace itk
@@ -186,6 +187,9 @@ namespace itk
StopConditionType m_StopCondition;
double m_StateOfConvergence;
unsigned long m_CurrentIteration;
+
+ /** Random number generator */
+ Statistics::MersenneTwisterRandomVariateGenerator::Pointer m_Generator;
/** Method to compute the learning rate at iteration k (a_k).*/
virtual double Compute_a( unsigned long k ) const;
diff --git a/Utilities/ITK/Code/Numerics/itkSingleValuedNonLinearVnlOptimizer.cxx b/Utilities/ITK/Code/Numerics/itkSingleValuedNonLinearVnlOptimizer.cxx
index fda071a951..88242f6d34 100644
--- a/Utilities/ITK/Code/Numerics/itkSingleValuedNonLinearVnlOptimizer.cxx
+++ b/Utilities/ITK/Code/Numerics/itkSingleValuedNonLinearVnlOptimizer.cxx
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkSingleValuedNonLinearVnlOptimizer.cxx,v $
Language: C++
- Date: $Date: 2005/07/26 16:17:41 $
- Version: $Revision: 1.14 $
+ Date: $Date: 2006/05/12 12:58:27 $
+ Version: $Revision: 1.15 $
Copyright (c) Insight Software Consortium. All rights reserved.
See ITKCopyright.txt or http://www.itk.org/HTML/Copyright.htm for details.
@@ -29,7 +29,8 @@ SingleValuedNonLinearVnlOptimizer
m_CostFunctionAdaptor = 0;
m_Maximize = false;
m_Command = CommandType::New();
- m_Command->SetCallbackFunction( this, &SingleValuedNonLinearVnlOptimizer::IterationReport );
+ m_Command->SetCallbackFunction( this,
+ &SingleValuedNonLinearVnlOptimizer::IterationReport );
m_CachedValue = 0;
m_CachedCurrentPosition.Fill(0);
m_CachedDerivative.Fill(0);
diff --git a/Utilities/ITK/Code/Numerics/itkSingleValuedVnlCostFunctionAdaptor.cxx b/Utilities/ITK/Code/Numerics/itkSingleValuedVnlCostFunctionAdaptor.cxx
index 22adb4c4f7..55237fd5e4 100644
--- a/Utilities/ITK/Code/Numerics/itkSingleValuedVnlCostFunctionAdaptor.cxx
+++ b/Utilities/ITK/Code/Numerics/itkSingleValuedVnlCostFunctionAdaptor.cxx
@@ -3,8 +3,8 @@
Program: Insight Segmentation & Registration Toolkit
Module: $RCSfile: itkSingleValuedVnlCostFunctionAdaptor.cxx,v $
Language: C++
- Date: $Date: 2005/03/30 22:26:05 $
- Version: $Revision: 1.18 $
+ Date: $Date: 2006/05/12 12:58:27 $
+ Version: $Revision: 1.19 $
Copyright (c) Insight Software Consortium. All rights reserved.
See ITKCopyright.txt or http://www.itk.org/HTML/Copyright.htm for details.
@@ -211,8 +211,8 @@ SingleValuedVnlCostFunctionAdaptor
return m_NegateCostFunction;
}
-/** Returns whether the cost function is going to be negated or not.
- * This is useful for adapting optimizers that are only minimizers. */
+/** This method reports iterations events. It is intended to
+ * help monitoring the progress of the optimization process. */
void
SingleValuedVnlCostFunctionAdaptor
::ReportIteration( const EventObject & event ) const
@@ -259,7 +259,6 @@ SingleValuedVnlCostFunctionAdaptor
-
} // end namespace itk
--
GitLab