enter_distributions_commands.m

Duda的《模式分类》第二版的配套的Matlab源代码

function enter_distributions_commands(command, param)

%This functions handles callbacks from the enter_distributions GUI

switch command
case 'change_parameter'
   %When there are changes to one of the parameters of an object, save it to file.
   %Used by the manual entry screen.
   
   load synthetic
   
   tag			= get(param, 'Tag');
   val 			= get(findobj('Tag', 'popNumber'), 'Value');
   zero_or_one = ~get(findobj('Tag', 'rbtClass0'), 'Value');
   num			= str2num(get(param, 'String'));
   
   switch tag
   case 'txtMeanX'
      if zero_or_one,
         distribution_parameters(2).mu(val,1) = num;
      else
         distribution_parameters(1).mu(val,1) = num;
      end
   case 'txtMeanY'
      if zero_or_one,
         distribution_parameters(2).mu(val,2) = num;
      else
         distribution_parameters(1).mu(val,2) = num;
      end
   case 'txtWeight'
      if zero_or_one,
         distribution_parameters(2).w(val) = num;
      else
         distribution_parameters(1).w(val) = num;
      end
   case 'txtCov11'
      if zero_or_one,
         distribution_parameters(2).sigma(val,1,1) = num;
      else
         distribution_parameters(1).sigma(val,1,1) = num;
      end
   case 'txtCov12'
      if zero_or_one,
         distribution_parameters(2).sigma(val,1,2) = num;
      else
         distribution_parameters(1).sigma(val,1,2) = num;
      end
   case 'txtCov21'
      if zero_or_one,
         distribution_parameters(2).sigma(val,2,1) = num;
      else
         distribution_parameters(1).sigma(val,2,1) = num;
      end
   case 'txtCov22'
      if zero_or_one,
         distribution_parameters(2).sigma(val,2,2) = num;
      else
         distribution_parameters(1).sigma(val,2,2) = num;
      end
   end
   
   save synthetic distribution_parameters
   
case 'change_class'
   %Change the display of a class, when one of the radio buttons is pressed.
   %Used by the manual entry screen.
   
   load synthetic
   
   h0 = findobj('Tag', 'rbtClass0');       
   h1 = findobj('Tag', 'rbtClass1');       
   h  = findobj('Tag', 'txtCount');
   hp = findobj('Tag', 'popNumber');                                                                     
   n0 = size(distribution_parameters(1).sigma,1);
   n1 = size(distribution_parameters(2).sigma,1);
   
   if param,
      set(h0, 'Value', not(get(h1, 'Value')));
   else
      set(h1, 'Value', not(get(h0, 'Value')));
   end
   
   
   %Set the number of objects (if they exist)
   if (get(h0, 'Value'))
      if (n0>0),
         set(h, 'String', ['There are ' num2str(n0) ' distributions in this class'])
         s = cell(n0, 1);                                                                                    
         for i=1:n0,                                                                                         
            s(i) =  cellstr(num2str(i));                                                                      
         end                                                                                                  
         set(hp, 'String', s')                                                                                 
         set(hp, 'Value', 1)                                                                                 
         set(hp, 'Max', n0);                                                                                  
      end
   else
      if (n1>0),
         set(h, 'String', ['There are ' num2str(n1) ' distributions in this class'])
         s = cell(n1, 1);                                                                                    
         for i=1:n1,                                                                                         
            s(i) =  cellstr(num2str(i));                                                                      
         end                                                                                                  
         set(hp, 'String', s')                                                                                 
         set(hp, 'Value', 1)                                                                                 
         set(hp, 'Max', n1);                                                                                     
      end
   end
   
   enter_distributions_commands('change_object')
   
case 'change_object'
   %When the user selects a different object, change the display to show that object.
   %Used by the manual entry screen.
   
   load synthetic
   
   h 	= findobj('Tag', 'popNumber');
   ht   = findobj('Tag', 'popType');
   hm1  = findobj('Tag', 'txtMeanX');
   hm2 	= findobj('Tag', 'txtMeanY');
   hw 	= findobj('Tag', 'txtWeight');
   hs11 = findobj('Tag', 'txtCov11');
   hs12 = findobj('Tag', 'txtCov12');
   hs21 = findobj('Tag', 'txtCov21');
   hs22 = findobj('Tag', 'txtCov22');
   hc   = findobj('Tag', 'txtCovCaption');
   
   val = get(h, 'Value');
   
   if get(findobj('Tag', 'rbtClass0'), 'Value')
      %Class 0 selected
      set(hm1,  'String', distribution_parameters(1).mu(val,1))
      set(hm2,  'String', distribution_parameters(1).mu(val,2))
      set(hs11, 'String', distribution_parameters(1).sigma(val,1,1))
      set(hs12, 'String', distribution_parameters(1).sigma(val,1,2))
      set(hs21, 'String', distribution_parameters(1).sigma(val,2,1))
      set(hs22, 'String', distribution_parameters(1).sigma(val,2,2))
      set(hw,   'String', distribution_parameters(1).w(val))
      if (strcmp(distribution_parameters(1).type(val), 'Gaussian'))
          set(ht,   'Value',  1);
          set(hs21, 'Visible', 'on');
          set(hs22, 'Visible', 'on');
          set(hc, 'String', 'Covariance of the Gaussian:'); 
      else
          set(ht,   'Value',  2);
          set(hs21, 'Visible', 'off');
          set(hs22, 'Visible', 'off');
          set(hc, 'String', 'Width of the distribution:'); 
      end
   else
      %Class 1 selected
      set(hm1,  'String', distribution_parameters(2).mu(val,1))
      set(hm2,  'String', distribution_parameters(2).mu(val,2))
      set(hs11, 'String', distribution_parameters(2).sigma(val,1,1))
      set(hs12, 'String', distribution_parameters(2).sigma(val,1,2))
      set(hs21, 'String', distribution_parameters(2).sigma(val,2,1))
      set(hs22, 'String', distribution_parameters(2).sigma(val,2,2))
      set(hw,   'String', distribution_parameters(2).w(val))
      if (strcmp(distribution_parameters(2).type(val), 'Gaussian'))
          set(ht,   'Value',  1);
          set(hs21, 'Visible', 'on');
          set(hs22, 'Visible', 'on');
      else
          set(ht,   'Value',  2);
          set(hs21, 'Visible', 'off');
          set(hs22, 'Visible', 'off');
      end
   end   
case 'change_type'
   val 			= get(findobj('Tag', 'popNumber'), 'Value');
   zero_or_one  = ~get(findobj('Tag', 'rbtClass0'), 'Value');
   ht           = findobj('Tag', 'popType');
   hs21         = findobj('Tag', 'txtCov21');
   hs22         = findobj('Tag', 'txtCov22');
   hc           = findobj('Tag', 'txtCovCaption');
   st           = get(ht, 'String');
   
   if (get(ht, 'Value') == 1)
       %Gaussian
       set(hc, 'String', 'Covariance of the Gaussian:'); 
       set(hs21, 'Visible', 'on');
       set(hs22, 'Visible', 'on');
   else
       %Uniform
       set(hc, 'String', 'Width of the distribution:'); 
       set(hs21, 'Visible', 'off');
       set(hs22, 'Visible', 'off');
   end
      
   load synthetic
   distribution_parameters(zero_or_one+1).type(val) = st(get(ht, 'Value'));     
   save synthetic distribution_parameters

   enter_distributions_commands('change_object')

otherwise
   error('Unknown command')
end