Evening lecture: 23
November 2004, at 18.15
Organizer: Australian
Statistical Society Inc.
Venue: Swinburne University of Technology, Hawthorn Campus, Melbourne, Australia
Topic:
Shape Understanding System: application of statistical
methods in designing of the system with the visual thinking capabilities
Zbigniew LES and Magdalena
LES
The
Queen Jadwiga Research Institute of Understanding
Australia
The aim of this presentation
is to illustrate the application of statistical methods in implementation
of the shape understanding system (SUS) and to present the application of visual inference in selected
areas of statistics. Visual inference and visual reasoning are part of the
visual thinking capabilities of the shape understanding system (SUS). Visual
reasoning involves transformation of the object description when passing
stages of the reasoning process. SUS is an example of the visual
understanding system where sensory information is transformed into the
multilevel representation in the concept formation process that is part of
the visual thinking capabilities. Understanding requires interpreting 2D
images as the real world objects, symbols and signs. Understanding is based
on a large number of highly varied abilities, called intelligence,
that can be measured. Abilities of SUS to understand the different categories of visual
objects were tested on the large classes of objects. It was shown that SUS was able to express the
visual perceptual data in the form of the linguistic expressions. SUS
performed visual reasoning based on the shape category and visual inference
based on combining the rules in which the visual concepts are embedded.
In this presentation the
focus is on the issues connected with shape, understanding, reasoning and
statistical methods. In the first part of the presentation the general
concepts such as understanding, visual thinking and reasoning will be
presented. The basic concepts of shape understanding method and selected
issues of the implementation of the shape understanding system (SUS) will
be introduced. In the second part, the statistical shape theory as well as
selected statistical methods applied in the pattern recognition and expert
systems such as Hidden Markov Models or Bayesian networks,
will be briefly discussed. The visual inference that is applied to solve
the problem of curve identification, a graphical investigation of the
characteristic points of the curve, visual tests, and identification of
statistical visual objects will be presented. Examples of SUS application
in statistics (the regression analysis, cluster analysis, discriminant analysis) will be also presented.
Back
Home
