Mark Benfield (LSU, USA)
Phil Culverhouse (Plymouth University, UK)
Invited Speaker:
Cabell Davis (Woods Hole Oceanographic Institution,
USA)
Advanced pattern recognition techniques are being
applied to plankton identification to automate sample specimen
counting to generic level. These tools, for example Zoo/Phyto Image
and Zooprocess/Plankton Identify offer
fast semi-automatic identification. They are free and can analyse the
output from a flatbed scanner (Zoo/phyto
Image), Zooscan and FlowCAM instruments as well as from digital cameras
and other sources. This workshop
will introduce the concepts and methods used, with some practical experience
in using the tools. Automation can
allow many thousands of specimens to be analysed daily. This workshop
is a must for those wishing to embrace
this new technology.
The workshop is split into 4 blocks. The first
block will present an overview of current practices (manual and
machine), together with the shortcomings of manual identification (if
you have a laptop, you can take part in an
identification experiment). We then introduce the basics of machine
identification (i.e., extracting measurements
from plankton images and using spreadsheets to show how images may be
grouped into clusters), using previously
prepared data from a Zooimage or a Zooscan machine. The third block
will cover issues of machine calibration
and using it in routine sample analysis. We will close with a look what
is happening in leading laboratories around
the world, and what the future holds.
