perClass Documentation
version 5.4 (7-Dec-2018)
SDSEGMENT  Segment image labels into connected components

   DATA=SDSEGMENT(DATA)

 INPUT
    DATA     SDDATA image created with SDIMAGE

 OUTPUT
    DATA     SDDATA image with object labels

 OPTIONS
    'minsize',N  Minimum object size, to be labeled (default: 10)
    'maxsize',M  Maximum number of regions (default: 5000)
    'remove'     Remove small regions (default: do not remove)

 DESCRIPTION
 SDSEGMENT finds connected components in the current labels of the image
 data set SDDATA (feature values are not used). It splits the current
 labels into connected regions and adds 'object' labels. Each class is
 processed separately. Regions smaller than 'minsize' are aggregated into
 one 'small object' category per class. With the 'remove' option, these
 small regions are removed from the returned data set.

 EXAMPLE
 Define conncted components by data.lab:
 >> a=sdsegment(data,'minsize',5)

 Get only regions larger than 5 pixels (select objects that are not
 small by regular expression):
 >> b=subset(a,'object','~/small')

 SEE ALSO
 SDIMAGE, SDEXTRACT