[mmdist] [Up] [mmopentransf] Image Transforms

mmgdist
Geodesic Distance Transform.

Synopsis

y = mmgdist( f, g, Bc = None, METRIC = None )

Implemented in Python.

Input

f Image Binary image.
g Image Binary image.

Marker image
Bc Structuring Element

(metric for distance).

Default: None (Diamond structuring element with radius 1)

METRIC String

'EUCLIDEAN' if specified.

Default: None

Output

y Image Gray-scale (uint8 or uint16) image.

uint16 (distance image).

Description

mmgdist creates the geodesic distance image y of the binary image f relative to the binary image g. The value of y at the pixel x is the length of the smallest path between x and f. The distances available are based on the Euclidean metrics and on metrics generated by a neighbourhood graph, that is characterized by a connectivity rule defined by the structuring element Bc. The connectivity for defining the paths is consistent with the metrics adopted to measure their length. In the case of the Euclidean distance, the space is considered continuos and, in the other cases, the connectivity is the one defined by Bc.

Examples

>>> f=mmbinary([
 [1,1,1,1,1,1],
 [1,1,1,0,0,1],
 [1,0,1,0,0,1],
 [1,0,1,1,0,0],
 [0,0,1,1,1,1],
 [0,0,0,1,1,1]])

              
>>> g=mmbinary([
 [0,0,0,0,0,0],
 [1,1,0,0,0,0],
 [0,0,0,0,0,0],
 [0,0,0,0,0,0],
 [0,0,0,0,0,0],
 [0,0,0,0,0,1]])

              

              
>>> y=mmgdist(f,g,mmsecross())

              

            

          

          

            

              
>>> f=mmreadgray('maze_bw.tif')

              

              
>>> g=mmintersec(f,0)

              
Warning: Converting input image from int32 to binary uint8.

              
>>> g=mmdrawv(g,uint16([[2],[2],[6],[6]]),uint16(1),'frect')

              

              
>>> y=mmgdist(f,g,mmsebox(),'EUCLIDEAN')

              

              
>>> mmshow(f,g)

              

              
>>> mmdtshow(y,200)

              

            

            

              
                

                  
                  
                  
                

                

                  f,g
                  y,200
                  
                

              
            

          

        

      

    

    

      
Equation

      

        

          
geodesic distance function using structuring element:

          

            
          

        

      

    

    

      
Limitations

      

        

                      To generate useful Distance transforms, the structuring elements must be symmetric and with the origin included.
          The Euclidean Distance transform is rounded to the nearest integer, since it is represented in an unsigned integer image. You should use the 
          mmsebox structuring element when computing the Euclidean Distance transform.
                  
        

      

    

    

      
Source Code

      

        
def mmgdist(f, g, Bc=None, METRIC=None):
    if Bc is None: Bc = mmsecross()
    assert METRIC is None,'Does not support EUCLIDEAN'
    fneg,gneg = mmneg(f),mmneg(g)
    y = mmgray(gneg,'uint16',1)
    ero = mmintersec(y,0)
    aux = y
    i = 1
    while not mmisequal(ero,aux):
        aux = ero
        ero = mmcero(gneg,fneg,Bc,i)
        y = mmaddm(y,mmgray(ero,'uint16',1))
        i = i + 1
    y = mmunion(y,mmgray(ero,'uint16'))
    return y
    

      

    

    

      
See also

      

        

          
            

              mmdist
              Distance transform.
            

            

              mmsecross
              Cross structuring element.
            

            

              mmsebox
              Create a box structuring element.
            

            

              mmfreedom
              Control automatic data type conversion.
            

            

              mmcero
              Erode an image conditionally.
            

          
        

      

    

    

      

        

          
                    [mmdist]
                
                    [Up]
                    
                    [mmopentransf]
                
          Python
        

        

          Copyright (c) 2003, Roberto A. Lotufo, UNICAMP-University of Campinas; Rubens C. Machado, CenPRA-Renato Archer Research Center.