How to set that image to 0 (false) at those locations, instead of plotting a marker over them? That would break it apart.

To divide that skeleton below : to a set of regions based on labels below : X=[1 0 1 ; 0 1 0 ; 1 0 1 ;]; Y=[1 0 1 ; 0 1 0 ; 1 0 0 ;]; O=[0 1 0 ; 1 0 1 ; 0 1 0 ;]; -=[0 0 0 ; 1 1 1 ; 0 0 0 ;]; L=[1 0 0 ; 1 0 0 ; 1 1 1 ;];

after detecting them using this code :

 n = 50 ;
 A = double( rand(n, n+1) > 0.5 ) ;
 B = 2*A - 1 ;
 patterns = {[1 0 1; 0 1 0; 1 0 1]; ... % X
             [1 0 1; 0 1 0; 1 0 0]; ... % Y
             [0 1 0; 1 0 1; 0 1 0]; ... % O
             [0 0 0; 1 1 1; 0 0 0]; ... % -
             [1 0 0; 1 0 0; 1 1 1]} ;   % L
 labels   = {'X', 'Y', 'O', '-', 'L'} ;
 matches  = cell( size(labels )) ;
 for pId = 1 : numel( patterns )
     nel    = numel( patterns{pId} ) ;
     ker    = 2*patterns{pId} - 1 ;    
     [r, c] = find( conv2(B, rot90(rot90(ker)), 'same') == nel ) ;
     matches{pId} = [r, c] ;
     figure(pId) ;  clf ;  hold on ;
     spy(A) ;
     plot( c, r, 'r+', 'MarkerSize', 20 ) ;
     title( sprintf( 'Matches for "%s" pattern', labels{pId} )) ;
     set( gcf, 'Units', 'normalized' ) ;
     set( gcf, 'Position', [0 0 0.4 0.7] ) ;
 end

How to set to 0 (false) at those locations, instead of plotting a marker over them? That would break it apart.

Any ideas, Help is much appreciated.

[sixwwwwww]

what do you mean by false here? do you want to place black spots there?

[mika]

Hello sixwwwwww,

Replacing 1 by 0 on the image below to break it apart,

 at those locations detected, instead of plotting a marker over them.

[mika]

Or can we break apart that image where every region contain a marker ?

help please !

thx

Answer 1

[sixwwwwww]

try this:

n = 50 ;
 A = double( rand(n, n+1) > 0.5 ) ;
 B = 2*A - 1 ;
 patterns = {[1 0 1; 0 1 0; 1 0 1]; ... % X
             [1 0 1; 0 1 0; 1 0 0]; ... % Y
             [0 1 0; 1 0 1; 0 1 0]; ... % O
             [0 0 0; 1 1 1; 0 0 0]; ... % -
             [1 0 0; 1 0 0; 1 1 1]} ;   % L
 labels   = {'X', 'Y', 'O', '-', 'L'} ;
 matches  = cell( size(labels )) ;
 for pId = 1 : numel( patterns )
     nel    = numel( patterns{pId} ) ;
     ker    = 2*patterns{pId} - 1 ;    
     [r, c] = find( conv2(B, rot90(rot90(ker)), 'same') == nel ) ;
     matches{pId} = [r, c] ;
     figure(pId) ;  clf ;  hold on ;
     spy(A) ;
     plot( c, r, 'r', 'LineWidth', 2, 'MarkerSize', 20 ) ;
     title( sprintf( 'Matches for "%s" pattern', labels{pId} )) ;
     set( gcf, 'Units', 'normalized' ) ;
     set( gcf, 'Position', [0 0 0.4 0.7] ) ;
 end

It will connect marker points and will show how different parts of image are separated

[mika]

Thank you for your response, But that's far from what i'm trying to do, can we break apart that image where every region contain a marker in an ordered way like what is shown by yellow dots(hand mad) for example in the image below?

there is the original image that i'm working on:

Help is much appreciated. thx

[sixwwwwww]

For this purpose we need more marker points. So how you would like to generate more marker points?

[mika]

The yellow points are not a markers i mean by drowing it with paint that this squeleton can be break apart there for example(breaking apart can mean replacing 1 by 0 )

Any ideas,

thx.

[sixwwwwww]

Following code going close to it but lines are crossing each other:

n = 50 ;
 A = double( rand(n, n+1) > 0.5 ) ;
 B = 2*A - 1 ;
 patterns = {[1 0 1; 0 1 0; 1 0 1]; ... % X
             [1 0 1; 0 1 0; 1 0 0]; ... % Y
             [0 1 0; 1 0 1; 0 1 0]; ... % O
             [0 0 0; 1 1 1; 0 0 0]; ... % -
             [1 0 0; 1 0 0; 1 1 1]} ;   % L
 labels   = {'X', 'Y', 'O', '-', 'L'} ;
 matches  = cell( size(labels )) ;
 for pId = 1 : numel( patterns )
     nel    = numel( patterns{pId} ) ;
     ker    = 2 * patterns{pId} - 1 ;    
     [r, c] = find( conv2(B, rot90(rot90(ker)), 'same') == nel ) ;
     matches{pId} = [r, c] ;
     spy(A) ;
     plot( c, r, 'k', 'LineWidth', 3, 'MarkerSize', 20 ), hold on
     set( gcf, 'Units', 'normalized' ) ;
     set( gcf, 'Position', [0 0 0.4 0.7] ) ;
 end

Maybe you try to find a way that they don't cross each other then you will get the figure as you need