Hi John,

thanks for the information about employment and the TV show (though I haven't
watched it yet, but it sounds fun!).

What is SRI? What I find most interesting about single-raster-scan connected
component labeling algorithms is how they handle the event of two independent
components connecting later in the course of receiving raster scan lines
(V-shaped objects for example). To my (somewhat limited) knowledge about
single-raster-scan image processing algorithms, the parameter extraction has to
be closely integrated with the component labeling which makes the whole design
somewhat unflexible.

Christian


--- In FPGA_CV@yahoogroups.com, John Kent <jekent@...> wrote:
>
> Hi Christian,
>
> I worked at the CSIRO in the mid 1990s. They developed an Area
> Parameter Accelerator that generated SRI parameters for the blobs. The
> connectivity was performed in one raster scan. The algorithm described
> in Haralick and Shapiro requires two passed, a forward pass and a
> reverse pass. All I needed from connectivity analysis is a perimeter
> point for each of the blobs.
>
> For ICT jobs in Victoria check out the Multimedia Victoria web site
> http://www.mmv.vic.gov.au/
> Look under "Find an ICT Company".
>
> If you are in Australia you might turn on the ABC at 8:30 pm ... There
> is a program "Rodney's Robots"
>
http://www.abc.net.au/tv/guide/netw/200909/programs/DO0728H001D2009-09-24T203000\
.htm
>
> John.
>
> Christian Gelinek wrote:
> >
> >
> > Hello everyone,
> >
> > my name is Christian and I have developed image processing systems
> > based on FPGAs during the last 6 years.
> >
> > > Has anyone tried implementing connectivity analysis as described in
> > > Haralick & Shapiro's Robot vision Vol 2 book ?
> > > I've looked at the pseudo code but it's a bit hard to decipher some of
> > > the details.
> >
> > I don't have this book, but connectivity analysis is an interesting
> > topic. Is the algorithm of the book designed to work in a single
> > raster scan pass?
> >
> > A former colleague of mine is (or should be) implementing an single
> > raster scan connectivity analysis algorithm for FPGAs at the moment.
> > But of course, with a single pass method, the output of this analysis
> > is not (can not be) the final result - imagine w-shaped objects
> > scanned from top to bottom. John, what kind of output from such an
> > algorithm are you looking for?
> >
> > And then I have another, rather off-topic question: does anyone of you
> > guys know about a company in Australia interested in employing an
> > FPGA/embedded software engineer like me?
> >
> > Regards
> > Christian
> > --
> > GRATIS für alle GMX-Mitglieder: Die maxdome Movie-FLAT!
> > Jetzt freischalten unter http://portal.gmx.net/de/go/maxdome01
> > <http://portal.gmx.net/de/go/maxdome01>
> >
>
> --
> http://www.johnkent.com.au
> http://members.optusnet.com.au/jekent
>