IEEE Robotics and Automation Society
                 Central New England Chapter

                  Tuesday, October 12, 2004
                 Informal Discussion: 6:00 PM
                       Program: 6:30 PM

      Color Machine Vision: Challenging Human Perception

                  Robert K. McConnell, Ph.D.
                        rkm@way2c.com
                    WAY-2C, Arlington, MA
                    http://www.way2c.com/


                           ABSTRACT

Fusing of images obtained in different spectral regions
and/or from different sensors can provide a large amount of
highly diagnostic information about objects in a scene. When
the fusion takes place in our optical systems we are said to
be seeing a "color" image.

Many people familiar with machine vision equate color
machine vision with location, identification, and/or
inspection of objects or organisms with simple color
distributions, simple boundaries, and simple backgrounds.
Few real applications meet these criteria. Conventional
wisdom suggests that addition of non-uniform lighting,
hardware pixillation effects and other sources of "noise"
will further complicate the situation.

Humans and other organisms, some with very limited natural
computational power, effortlessly incorporate information
from complex color distributions into their decision-making
processes. This strongly suggests that traditional
multispectral image interpretation methods may be using less
than optimum approaches.

The speaker will discuss and demonstrate WAY-2C, an
unconventional machine vision and image interpretation
system. The system has been successfully deployed in a
variety of challenging real-world problems including
assembly inspection, color based sorting, and remote sensing
image interpretation. If time permits human vs. machine
color-based recognition experiments will be performed on
randomly chosen samples provided by members of the audience.

Test images can be seen on the web at the following URLs:

	http://www.way2c.com/puzzlebw.jpg
	http://www.way2c.com/puzzle.jpg


                      SPEAKER BIOGRAPHY

Dr. McConnell received a B.Sc.E. in Geological Engineering
from Princeton University and an M.A.Sc. in Applied Geology
and Ph.D. in Physics (Geophysics) from the University of
Toronto. This background provided him with image
interpretation experience ranging from microscopy to remote
sensing and mathematical tools to help automate such
interpretation. He has held a variety of positions including
member of the technical staff at Arthur D. Little, Inc.;
Visiting Professor of Oceanography, at the University of
Rhode Island; and president of Earth Sciences Research Inc.
More recently he was a founder and chairman of Wayland
Research Inc.

Recent responsibilities include applications of minimum
description analysis to the analysis of color,
hyperspectral, and other images. They include development of
WAY-2C, a color machine vision and image interpretation
system used around the world in such industries as
automotive, food, recycling, and high level nuclear waste
disposal.

Dr. McConnell is the author of over 35 professional papers
in the fields of geophysics and image analysis and holder of
a comprehensive patent on pattern recognition. He is a
Senior Member of the Society of Manufacturing Engineers,
active on the steering committee of the Central New England
Chapter of the IEEE Robotics and Automation Society, and a
member of the American Geophysical Union and the AAAS.


                     MEETING INFORMATION

The IEEE Robotics and Automation Society will meet on
Tuesday, October 12, 2004, at Wellesley High School at
6:00 PM for informal discussions and for the formal
presentation between 6:30 and 7:30 PM. The group will have a
no-host dinner afterwards at Bertucci's, where more
conversations can take place with the guest speaker. The
meetings are open to the general public, and all are welcome
at the dinner afterwards.

For more information, contact Rich Maynard as follows:

     phone:	978-439-5511 ext. 6221
     email:	chair@robotics-boston.org


                          DIRECTIONS

 From Route 128, take either Route 9 or Route 16 west. From
the junction of Routes 9 and 16, follow Route 16 (Washington
Street) west past the Wellesley Hills commuter rail station
(within walking distance of the school), then turn left onto
Rice Street. Wellesley High School is on the left. Parking
is available in a lot just beyond the school.

For a web map showing the location of Wellesley High School,
go to the Yahoo map site at http://maps.yahoo.com/ and enter
the following information in the indicated locations:

	Address:		50 Rice St
	City, State or ZIP:	Wellesley, MA

The star marker is misplaced slightly to the south. Imagine
it about 1 cm toward the top of the map at Rice St.

The GPS coordinates for Rice St in front of the entrance to
the high school building are approximately as follows:

	N 42 18.137'	W 71 16.811'

The GPS coordinates for the parking lot in front of the
entrance to Bertucci's are approximately as follows:

	N 42 18.492'	W 71 16.691'

For more information about our Section and Chapter, visit
our web site at our registered domain URL

               http://www.robotics-boston.org/