nic_spm_bwlabel   

 Connected component labelling in 2D or 3D. 

 FORMAT: [L,NUM] = spm_bwlabel(bw,conn)
 
 The calling interface has been modelled on the 
 image processing toolbox routine bwlabel. See 
 the help for that routine for further information.

 Input:
 bw:         : (Binary) image to perform labelling on. Can
               be 2 or 3D.
 conn        : Connectivity criterion. Could be 6(surface)
               18(edge) or 26(corner). For a 2D bw these
               correspond to 4, 8 and 8 respectively.

 Output:
 L           : Connected component image, i.e. image where
               each non-zero voxel in bw will have a value
               corresponding to its label.
 NUM         : Number of connected components in L.


 The implementation is not recursive (i.e. will no crash for
 large connected components) and is losely based on
 Thurfjell et al. 1992, A new three-dimensional connected
 components labeling algorithm with simultaneous object
 feature extraction capability. CVGIP: Graphical Models 
 and Image Processing 54(4):357-364.
_______________________________________________________________________
 Copyright (C) 2005 Wellcome Department of Imaging Neuroscience

0001 function varargout = nic_spm_bwlabel(varargin)
0002 % Connected component labelling in 2D or 3D.
0003 %
0004 % FORMAT: [L,NUM] = spm_bwlabel(bw,conn)
0005 %
0006 % The calling interface has been modelled on the
0007 % image processing toolbox routine bwlabel. See
0008 % the help for that routine for further information.
0009 %
0010 % Input:
0011 % bw:         : (Binary) image to perform labelling on. Can
0012 %               be 2 or 3D.
0013 % conn        : Connectivity criterion. Could be 6(surface)
0014 %               18(edge) or 26(corner). For a 2D bw these
0015 %               correspond to 4, 8 and 8 respectively.
0016 %
0017 % Output:
0018 % L           : Connected component image, i.e. image where
0019 %               each non-zero voxel in bw will have a value
0020 %               corresponding to its label.
0021 % NUM         : Number of connected components in L.
0022 %
0023 %
0024 % The implementation is not recursive (i.e. will no crash for
0025 % large connected components) and is losely based on
0026 % Thurfjell et al. 1992, A new three-dimensional connected
0027 % components labeling algorithm with simultaneous object
0028 % feature extraction capability. CVGIP: Graphical Models
0029 % and Image Processing 54(4):357-364.
0030 %_______________________________________________________________________
0031 % Copyright (C) 2005 Wellcome Department of Imaging Neuroscience
0032 
0033 % Jesper Andersson
0034 % $Id: spm_bwlabel.m 159 2005-05-16 14:00:56Z guillaume $
0035 
0036 
0037 error('spm_bwlabel.c not compiled - see Makefile');
0038 
0039 
0040 
0041