00001 /*!@file Learn/SOFM.H Self-Organizing Map neural network */ 00002 00003 // //////////////////////////////////////////////////////////////////// // 00004 // The iLab Neuromorphic Vision C++ Toolkit - Copyright (C) 2000-2005 // 00005 // by the University of Southern California (USC) and the iLab at USC. // 00006 // See http://iLab.usc.edu for information about this project. // 00007 // //////////////////////////////////////////////////////////////////// // 00008 // Major portions of the iLab Neuromorphic Vision Toolkit are protected // 00009 // under the U.S. patent ``Computation of Intrinsic Perceptual Saliency // 00010 // in Visual Environments, and Applications'' by Christof Koch and // 00011 // Laurent Itti, California Institute of Technology, 2001 (patent // 00012 // pending; application number 09/912,225 filed July 23, 2001; see // 00013 // http://pair.uspto.gov/cgi-bin/final/home.pl for current status). // 00014 // //////////////////////////////////////////////////////////////////// // 00015 // This file is part of the iLab Neuromorphic Vision C++ Toolkit. // 00016 // // 00017 // The iLab Neuromorphic Vision C++ Toolkit is free software; you can // 00018 // redistribute it and/or modify it under the terms of the GNU General // 00019 // Public License as published by the Free Software Foundation; either // 00020 // version 2 of the License, or (at your option) any later version. // 00021 // // 00022 // The iLab Neuromorphic Vision C++ Toolkit is distributed in the hope // 00023 // that it will be useful, but WITHOUT ANY WARRANTY; without even the // 00024 // implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR // 00025 // PURPOSE. See the GNU General Public License for more details. // 00026 // // 00027 // You should have received a copy of the GNU General Public License // 00028 // along with the iLab Neuromorphic Vision C++ Toolkit; if not, write // 00029 // to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, // 00030 // Boston, MA 02111-1307 USA. // 00031 // //////////////////////////////////////////////////////////////////// // 00032 // 00033 // Primary maintainer for this file: Lior Elazary <elazary@usc.edu> 00034 // $HeadURL: svn://isvn.usc.edu/software/invt/trunk/saliency/src/Learn/SOFM.H $ 00035 // $Id: SOFM.H 13223 2010-04-15 00:52:46Z lior $ 00036 // 00037 00038 #ifndef LEARN_SOFM_H_DEFINED 00039 #define LEARN_SOFM_H_DEFINED 00040 00041 #include "Image/Image.H" 00042 #include "Image/Pixels.H" 00043 #include "Image/ColorOps.H" 00044 00045 class SOFM 00046 { 00047 public: 00048 enum DISTANCE_MEASURE {EUCLIDEAN, KL,L2GMM}; 00049 00050 SOFM(const char *netname, int InputSize, int x, int y); //init with map size x y 00051 ~SOFM(); 00052 void SetLearningRate(unsigned long long learning_time); 00053 void Train(float *Input, float *Target, double score); 00054 void organize(std::vector<double> &input); 00055 void organize(const Image<float>& input); 00056 Image<float> getMap(); 00057 Image<float> getActMap(); 00058 Image<PixRGB<byte> > getWeightsImage(); 00059 std::vector<float> getWeights(const Point2D<int> loc); 00060 00061 void ReadNet(const char *filename); 00062 void WriteNet(const char *filename); 00063 double RandomRange(double Low, double High); 00064 double RandomBinaryRange(double Low, double High); 00065 double Neighborhood(int i); 00066 void RandomWeights(int min = 0, int max = 255); 00067 void ZeroWeights(); 00068 void SetInput(float *in); 00069 void setInput(const std::vector<double> &in); 00070 void setInput(const Image<float> &in); 00071 void Propagate(DISTANCE_MEASURE dm = EUCLIDEAN); 00072 Point2D<int> getWinner(double& val); 00073 00074 00075 //private: 00076 struct Layer { 00077 int Units; //number of units 00078 double *Output; //output of i unit 00079 float **Weight; //connection weightto i unit 00080 double *StepSizeX; //size of search steps of i unit 00081 double *StepSizeY; //size of search steps of i unit 00082 double *dScoreMean; //mean score delta of i unit 00083 double *score; // the previous score for this move 00084 int *BadDim; // the dimintion that we need to work on 00085 double *Lambda; //for displyaing the neigborhood of units 00086 }; 00087 00088 Layer *InputLayer; 00089 Layer *KohonenLayer; 00090 int itsWinner; 00091 double itsWinnerValue; 00092 int itsLooser; 00093 double itsLooserValue; 00094 int MapSizeX; 00095 int MapSizeY; 00096 00097 const char *name; // the name of the network for saving 00098 double KohonenAlpha; //learning rate of Kohonen layer 00099 double OutAlpha; //learning rate of output layer 00100 double StepAlpha; //learning rate for step sizes 00101 double Gamma; //smoothing factor for score deltas 00102 double Sigma; //width of neighoborhood 00103 unsigned long long itsLearningTime; //The current learning time 00104 00105 00106 void InitSofm(); 00107 00108 00109 }; 00110 00111 // ###################################################################### 00112 /* So things look consistent in everyone's emacs... */ 00113 /* Local Variables: */ 00114 /* indent-tabs-mode: nil */ 00115 /* End: */ 00116 00117 #endif // LEARN_SOFM_H_DEFINED
1.6.3