/*========================================================================= * * Copyright Insight Software Consortium * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0.txt * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. * *=========================================================================*/ #ifndef itkKullbackLeiblerCompareHistogramImageToImageMetric_h #define itkKullbackLeiblerCompareHistogramImageToImageMetric_h #include "itkCompareHistogramImageToImageMetric.h" namespace itk { /** \class KullbackLeiblerCompareHistogramImageToImageMetric * \brief Computes the Kubler Lieblach(KL) metric between the histogram * of the two images to be registered and a training histogram. * * This class is templated over the type of the fixed and moving * images to be compared. * * This class computers the KL-metric by comparing the histograms * of the testing histogram formed by the overlap of intensities in * the images, to a training histogram. It is based on the * following paper: * * Albert C.S. Chung, William M. Wells III, Alexander Norbash, and * W. Eric L. Grimson, Multi-modal Image Registration by * Minimising Kullback-Leibler Distance, In Medical Image Computing * and Computer-Assisted Intervention - MICCAI 2002, LNCS 2489, * pp. 525 - 532. * * The metric is given by KL(P_test||P_train) * = Sum_{i1,i2} P_test(i1,i2) std::log(P_test(i1,i2)/P_train(i1,i2)) * where P_test and P_train are probabilities given my normalized * histograms, and i1 and i2 are the intensity bins in the histogram. * * \par PARAMETERS * Epsilon is added to every bin in both histograms. This prevents * division by zero problems. Epsilon should generally be set to a * number smaller than one divided by the total number bins in * the histogram. So, for a 256 by 256 histogram, Epsilon should be * much less than 1e-5. Tests have shown that choices of epsilon are * not very important as long as it is small enough. The default is 1e-12. * I doubt you will need to change it. * * \author Samson Timoner * * \par SUPPORT * This work was supported by the Functional Imaging Research in * Schizophrenia Testbed (FIRST) Biomedical Informatics Research * Network (BIRN, http://www.birncommunity.org/), which is funded by the * National Center for Research Resources at the National * Institutes of Health (NIH). This work is also funded by the * Neuroimage Analysis Center (P41 RR13218). * * \ingroup RegistrationMetrics * \ingroup ITKRegistrationCommon */ template< typename TFixedImage, typename TMovingImage > class ITK_TEMPLATE_EXPORT KullbackLeiblerCompareHistogramImageToImageMetric: public CompareHistogramImageToImageMetric< TFixedImage, TMovingImage > { public: /** Standard class typedefs. */ typedef KullbackLeiblerCompareHistogramImageToImageMetric Self; typedef CompareHistogramImageToImageMetric< TFixedImage, TMovingImage > Superclass; typedef SmartPointer< Self > Pointer; typedef SmartPointer< const Self > ConstPointer; /** Method for creation through the object factory. */ itkNewMacro(Self); /** Run-time type information (and related methods). */ itkTypeMacro(KullbackLeiblerCompareHistogramImageToImageMetric, HistogramImageToImageMetric); /** Types transferred from the base class */ typedef typename Superclass::RealType RealType; typedef typename Superclass::TransformType TransformType; typedef typename Superclass::TransformPointer TransformPointer; typedef typename Superclass::ConstPointer TransformConstPointer; typedef typename Superclass::TransformParametersType TransformParametersType; typedef typename Superclass::TransformJacobianType TransformJacobianType; typedef typename Superclass::GradientPixelType GradientPixelType; typedef typename Superclass::MeasureType MeasureType; typedef typename Superclass::DerivativeType DerivativeType; typedef typename Superclass::FixedImageType FixedImageType; typedef typename Superclass::MovingImageType MovingImageType; typedef typename Superclass::FixedImageConstPointer FixedImageConstPointer; typedef typename Superclass::MovingImageConstPointer MovingImageConstPointer; typedef typename Superclass::HistogramType HistogramType; typedef typename Superclass::HistogramSizeType HistogramSizeType; typedef typename Superclass::MeasurementVectorType HistogramMeasurementVectorType; typedef typename Superclass::HistogramFrequencyType HistogramFrequencyType; typedef typename Superclass::HistogramIteratorType HistogramIteratorType; typedef typename Superclass::HistogramPointerType HistogramPointerType; typedef typename Superclass::InterpolatorType InterpolatorType; typedef typename Superclass::InterpolatorPointer InterpolatorPointer; /** Set epsilon, which is added to each bin in both Histogram */ itkSetMacro(Epsilon, double); /** Get epsilon, the histogram frequency to use if the frequency is 0 */ itkGetConstReferenceMacro(Epsilon, double); /** Return the number of parameters required by the Transform */ unsigned int GetNumberOfParameters(void) const ITK_OVERRIDE { return this->GetTransform()->GetNumberOfParameters(); } /** Forms the histogram of the training images to prepare to evaluate the */ /** metric. Must set all parameters first */ void Initialize() ITK_OVERRIDE; protected: /** Constructor is protected to ensure that \c New() function is used to create instances. */ KullbackLeiblerCompareHistogramImageToImageMetric(); virtual ~KullbackLeiblerCompareHistogramImageToImageMetric() ITK_OVERRIDE {} void PrintSelf(std::ostream & os, Indent indent) const ITK_OVERRIDE; /** Form the Histogram for the Training data */ void FormTrainingHistogram(); /** Evaluates the mutual information from the histogram. */ virtual MeasureType EvaluateMeasure(HistogramType & histogram) const ITK_OVERRIDE; double m_Epsilon; private: // Purposely not implemented. KullbackLeiblerCompareHistogramImageToImageMetric(Self const &); void operator=(Self const &); // Purposely not implemented. }; } // End namespace itk. #ifndef ITK_MANUAL_INSTANTIATION #include "itkKullbackLeiblerCompareHistogramImageToImageMetric.hxx" #endif #endif // itkKullbackLeiblerCompareHistogramImageToImageMetric_h