ENH: Simplify the histogram summary

no need for the max frequencies; only need to know if they are different

Modules/Fusion/MajorityVoting/include/otbNeighborhoodMajorityVotingImageFilter.h

@@ -175,10 +175,8 @@ protected:
   struct HistoSummary
   {
     unsigned int freqCenterLabel;
-    unsigned int majorityFreq;
     PixelType majorityLabel;
-    unsigned int secondFreq;
-    PixelType secondLabel;
+    bool majorityUnique;
   };
 
   struct CompareHistoFequencies

Modules/Fusion/MajorityVoting/include/otbNeighborhoodMajorityVotingImageFilter.txx

@@ -72,8 +72,7 @@ NeighborhoodMajorityVotingImageFilter<TInputImage,
     else
       {
       //If the majorityLabel is NOT unique in the neighborhood
-      if(histoSummary.secondFreq == histoSummary.majorityFreq && 
-         histoSummary.secondLabel != histoSummary.majorityLabel)
+      if(!histoSummary.majorityUnique)
         {
         if (m_KeepOriginalLabelBool == true)
           {
@@ -123,10 +122,8 @@ NeighborhoodMajorityVotingImageFilter<TInputImage, TOutputImage,
   typename NeighborhoodMajorityVotingImageFilter<TInputImage, TOutputImage,
                                                  TKernel>::HistoSummary result;
   result.freqCenterLabel = histoNeigh[nit.GetCenterPixel()];
-  result.majorityFreq = histoNeighVec[0].second;
   result.majorityLabel = histoNeighVec[0].first;
-  result.secondFreq = histoNeighVec[1].second;
-  result.secondLabel = histoNeighVec[1].first;
+  result.majorityUnique = (histoNeighVec[0].second != histoNeighVec[1].second);
   return result;
 }