MORE THAN 25 years
experience developing image processing and pattern recognition systems.
Created new learning, information extraction, and classification
techniques for biomedical, remote sensing and object recognition
applications. Skilled in algorithm development, software generation,
smart controls and vision system design. Received a Ph.D. in Electrical
Engineering in 1977 from Carnegie Mellon University. Doctoral work was
in biomedical pattern recognition and information theory. Published many
technical articles, reports and conference papers on image analysis,
object recognition, and learning and classification techniques.
Biomedical technology – generated a novel
information tree algorithm for the analysis of dermatoglyphic (finger
and hand print) data as a diagnostic aid in determination of some
genetic disorders. Developed a statistical package for the use of
medical researchers. Applied clustering and classification techniques
to a patient history database for evaluation of the effectiveness of
alternative treatments. Created automated cell screening algorithms
which select significant phenomena for later rapid review and
classification by experts. Designed a fully automated microscope
vision system for analysis of blood samples. This system is able to
automatically focus, color correct, find red blood cells, locate and
map white blood cells, and provide a six part differential analysis.
Multi-sensor systems – designed data fusion
techniques for combining information from multiple sensors for
automated object recognition. Created advanced graph-matching
recognition algorithms capable of classifying objects even under
conditions of partial occlusion. Extensive research with visible
light, infrared, real beam radar and synthetic aperture radar (SAR)
based processing. Applied classical statistical pattern recognition
and neural net techniques to image data for enhanced feature
extraction and recognition. Developed advanced automated target
recognition (ATR) algorithms for recognition of partially occluded
objects. Analyzed and specified stereo collection systems for tallying
and measuring parts and products.
Remote sensing – developed techniques for
extracting information from satellite data and automatically
generating a ground feature database. Utilized stereo extraction
techniques for automated terrain elevation map creation. Refined
techniques for combining multi-look imagery to enhance object
signatures and extract information.
Founded RSI in 1994:
RSI is a small company which develops new intelligent
software products and provides consulting support for vision system
development and object recognition applications. RSI developed the
Recognition Toolkit, a software package for building image-based object
recognition products. The Recognition Toolkit is an add-on product to
the KBVision and Aphelion image processing software of Amerinex Applied
Imaging. The Recognition Toolkit contains evaluation tools, a set of
classifiers, and a corresponding set of learning and training routines
to permit rapid development of tailored recognition software.
Biomedical Information Processing Coordinator for a
joint program between Carnegie Mellon University and the A. I. duPont
Institute. Project leader on many recognition and image analysis efforts
at The Analytic Sciences Corporation and group leader for algorithms and
modeling in the research laboratory at Textron Systems Corporation.