Hi,

We have recently developed an iterative algorithm for automated
segmentation of image stacks.  Our method out performs the standard
segmentation (simple threshold) method and is able to extract objects of
different intensities in the same plane, individual objects that contain a
range of intensities and can also separate 'touching' objects.  However, we
have been criticised for not comparing our algorithm with state-of-the-art
segmenters.  There is a load of literature in the pattern recognition and
computer vision journals but it is difficult to find/implement the source
code in order to make a comparison.  Anyway, there certainly isnt very much
that is designed to segment confocal data sets.

I would be interested to hear how other people in this group do their
segmentation.  Does anyone know of any routines that are semi-automated,
iterative and work well with confocal data?.

cheers

Craig.


______________________________________________________________________________
Craig J. Daly
Research Associate                     tel:       0141 330 6606
CRI Heart Failure                      fax:       0141 330 4100
West Medical Building                  email:     [log in to unmask]
University of Glasgow
Glasgow G12 8QQ.

Web Site                               http://130.209.48.61/pages/vascular.htm
____________________________________________________________________________
__