ml(rtees): use theRNG()

alalek · alalek · commit 526f62a90a60 · 2019-12-28T22:09:42.000Z
diff --git a/modules/ml/include/opencv2/ml.hpp b/modules/ml/include/opencv2/ml.hpp
@@ -1276,9 +1276,6 @@ class CV_EXPORTS_W RTrees : public DTrees
     /** @copybrief getTermCriteria @see getTermCriteria */
     CV_WRAP virtual void setTermCriteria(const TermCriteria &val) = 0;
 
-    /** Set initial random number generator state for training the forest. */
-    virtual void setForestTrainRNG(const RNG& rng) = 0;
-
     /** Returns the variable importance array.
     The method returns the variable importance vector, computed at the training stage when
     CalculateVarImportance is set to true. If this flag was set to false, the empty matrix is
diff --git a/modules/ml/src/rtrees.cpp b/modules/ml/src/rtrees.cpp
@@ -90,12 +90,12 @@ class DTreesImplForRTrees CV_FINAL : public DTreesImpl
         CV_TRACE_FUNCTION();
         DTreesImpl::clear();
         oobError = 0.;
-        rng = initRng;
     }
 
     const vector<int>& getActiveVars() CV_OVERRIDE
     {
         CV_TRACE_FUNCTION();
+        RNG &rng = theRNG();
         int i, nvars = (int)allVars.size(), m = (int)activeVars.size();
         for( i = 0; i < nvars; i++ )
         {
@@ -134,6 +134,7 @@ class DTreesImplForRTrees CV_FINAL : public DTreesImpl
     bool train( const Ptr<TrainData>& trainData, int flags ) CV_OVERRIDE
     {
         CV_TRACE_FUNCTION();
+        RNG &rng = theRNG();
         CV_Assert(!trainData.empty());
         startTraining(trainData, flags);
         int treeidx, ntrees = (rparams.termCrit.type & TermCriteria::COUNT) != 0 ?
@@ -424,8 +425,6 @@ class DTreesImplForRTrees CV_FINAL : public DTreesImpl
     double oobError;
     vector<float> varImportance;
     vector<int> allVars, activeVars;
-    RNG rng;
-    RNG initRng = RNG((uint64) - 1);
 };
 
 
