SPEAKER: DAVID M. MCKEOWN
Principal Research Computer Scientist and Head,
ABSTRACT:
Computer vision's fundamental problem in all of these fields is one of
"signal-to-symbol" transformation, going from a 2D array of "pixels"
to a description of image content or 3D scene semantics. Thirty years
ago, Raj Reddy and his students designed and implemented the first
primitive face recognition programs based on simple symbolic
descriptions of the human facial anatomy.
From this work, a sequence of image interpretation systems in various
application domains were developed under Raj's model of the importance
of semantics, domain knowledge, and the need for system-based
performance evaluation. Raj's early hand in the formulation of the
"CMU approach" to computer vision has allowed this area to grow
into many major efforts within the School of Computer Science.
In this talk I would like to reflect on some early work with Raj
on geospatial databases and observe how the CMU approach to this
problem domain has evolved over 25 years into automated systems for
the rapid construction and visualization of geospatial data.
SPEAKER BIO:
His research interests in computer science are in the areas of image
understanding for the automated analysis of remotely sensed imagery,
digital mapping and image/map database systems, simulation of
large scale geospatial environments, and artificial intelligence.
Mr. McKeown is a member of ACM, IEEE, AAAI, the American Society for
Photogrammetry and Remote Sensing (ASPRS), and Sigma Xi. He is chair of
Working Group II/6 on Integration of Image Understanding into
Cartographic Systems (1996-2000) and was co-chair of Intercommission
Working Group II/III on Digital Photogrammetric Systems of the
International Society for Photogrammetry and Remote Sensing (ISPRS) from
1992-1996.
Significant external awards include the 1995 Photogrammetric Award
(Fairchild) by ASPRS "in recognition of his sustained contributions
in computer science research appled to digital mapping". He was
presented with the 1996 Schermerhorn Award by the Netherlands Society
of Earth Observation and Geo-Informatics and ISPRS for his work
"in promoting scientific research and technical exchange meetings in
the area of computer vision and photogrammetry".
Digital Mapping Laboratory, Computer Science Department, Carnegie Mellon University
Image Processing to Image Understanding: Signal to Symbols
to Visualization
Over the last thirty years computer vision research has broadened
from concerns of individual image acquisition and iconic processing
to the automated extraction of useful information from streams of
imagery collected by a variety of sensors. As a scientific endeavor,
computer vision research has greatly matured and spun off many diverse
application areas in the fields of medicine, robotics, environmental
remote sensing, mapping and geographic information systems, and
entertainment.
David M. McKeown, Jr. is a Principal Research Computer Scientist
in the School of Computer Science at Carnegie Mellon University and
has been a member of the research faculty since 1986. Prior to
joining the faculty he was a researcher at Carnegie Mellon from
1975, a Research Associate at George Washington University and a
Member of the Technical Staff at NASA Goddard Space Flight Center,
Greenbelt, Maryland (1974--1975). Upon his arrival at CMU he joined
the HEARSAY II Speech Understanding Project and subsequently became
one of the initial researchers on the newly created ARPA Image
Understanding Program, both organized by Professor Reddy. He is
currently the head of the Digital Mapping Laboratory within the
Computer Science Department.