Beta behavior over SOM training phase.
This class implements an evolution of the weightening coefficient over the SOM training. It is issued from A. Cziho's PhD: "Compression d'images et analyse de contenu par quantification vectorielle" PhD dissertation, University of Rennes I, Rennes, France. May 5th, 1999.
Its behavior is decomposed into two steps depending on the number of iterations:
where stands for IterationThreshold.
CzihoSOMLearningBehaviorFunctor uses some parameters of the SOM class such as: BetaInit, BetaEnd, NumberOfIterations, but also NeighborhoodSizeInit which may be (surprisingly) required for the IterationThreshold.
The functor function uses
parameters, that is why it is necessary to call a specific method for
- See Also
Definition at line 60 of file otbCzihoSOMLearningBehaviorFunctor.h.