Susan Erickson wrote:

This was posted on the Orchid Digest. It should carry the news
source from a London paper.


Quote from OD --
From a news story : "LONDON, England...

A group of London-based scientists hope computer software they
are creating will help uncover previously unknown plant ...
species.

[snip]

The software program works by combining artificial intelligence
and computer vision technologies, which will load computers with
virtual collections of identified specimens.
...
Hi Sue,

Thanks. This was interesting. I have visited cnn.com where it was
originally published, and will print it out later. I really miss
working in a university context where I'd have the freedom and
resources to pursue such an interesting project!

Combining AI with computer vision? Talk about overkill! And this is
overkill that will take years to implement and probably decades before
it can be relied on! And I'd expect that it will, as conceived, be
outrageously expensive. It is a great research project and I'd love to
have the opportunity to work on it, but to be honest, if the purpose is
to identify plants, animals, fungae or any other taxon, one needs only
some basic software technologies that have existed for well over a
decade, and a few that are even older. What I have conceived could, if
I had funding, be implemented in less than a year, with another couple
years to build up the database. And at that point, it would be as
reliable as any taxonomist (since it would be doing automatically what
taxonomists, or at least numerical taxonomists, normally do), and more
useful than most since it would provide useful results for any kind of
organism while taxonomists are normally specialized on a small group of
taxa.

And I could make it more useful to research scientists eventually by
adding support for DNA finger printing of specimens and code to analyse
DNA sequence data; something that may prove useful in assessing
relationships among horticultural treasures. Imagine the professional
potential fr breeders and other vendors if they can have their plants
finger printed. If enough of them do so, relationships among breeding
lines, or plants of unknown provenance, could be assessed.

Thanks again.

Cheers,

Ted



R.E. (Ted) Byers, Ph.D., Ed.D.
R & D Decision Support Solutions
http://www.randddecisionsupportsolutions.com/
Healthy Living Through Informed Decision Making