ENH: remove TrainNormalizer

Modules/Applications/AppClassification/include/otbTrainSharkKMeans.hxx

@@ -23,6 +23,7 @@
 #include "otbLearningApplicationBase.h"
 #include "otbSharkKMeansMachineLearningModel.h"
 #include "otbStatisticsXMLFileReader.h"
+#include "shark/Models/Normalizer.h"
 
 namespace otb
 {

Modules/Learning/Unsupervised/include/otbSharkKMeansMachineLearningModel.h

@@ -48,7 +48,6 @@
 #include "shark/Models/Clustering/Centroids.h"
 #include "shark/Models/Clustering/ClusteringModel.h"
 #include "shark/Algorithms/KMeans.h"
-#include "shark/Algorithms/Trainers/NormalizeComponentsUnitVariance.h"
 
 #if defined(__GNUC__) || defined(__clang__)
 #pragma GCC diagnostic pop
@@ -146,9 +145,6 @@ protected:
   virtual void DoPredictBatch(const InputListSampleType *, const unsigned int &startIndex, const unsigned int &size,
                               TargetListSampleType *, ConfidenceListSampleType * = nullptr, ProbaListSampleType * = nullptr) const override;
 
-  template<typename DataType>
-  shark::Normalizer<> TrainNormalizer(const DataType &data) const;
-
   /** PrintSelf method */
   void PrintSelf(std::ostream &os, itk::Indent indent) const override;
 

Modules/Learning/Unsupervised/include/otbSharkKMeansMachineLearningModel.hxx

@@ -35,7 +35,6 @@
 
 #include "otb_shark.h"
 #include "otbSharkUtils.h"
-#include "shark/Algorithms/Trainers/NormalizeComponentsUnitVariance.h" //normalize
 #include "shark/Algorithms/KMeans.h" //k-means algorithm
 #include "shark/Models/Clustering/HardClusteringModel.h"
 #include "shark/Models/Clustering/SoftClusteringModel.h"
@@ -82,18 +81,6 @@ SharkKMeansMachineLearningModel<TInputValue, TOutputValue>
   m_ClusteringModel = boost::make_shared<ClusteringModelType>( &m_Centroids );
 }
 
-template<class TInputValue, class TOutputValue>
-template<typename DataType>
-shark::Normalizer<>
-SharkKMeansMachineLearningModel<TInputValue, TOutputValue>
-::TrainNormalizer(const DataType &data) const
-{
-  shark::Normalizer<> normalizer;
-  shark::NormalizeComponentsUnitVariance<> normalizingTrainer( true );//zero mean
-  normalizingTrainer.train( normalizer, data );
-  return normalizer;
-}
-
 template<class TInputValue, class TOutputValue>
 typename SharkKMeansMachineLearningModel<TInputValue, TOutputValue>
 ::TargetSampleType