BMVA
British Machine Vision Association and Society for Pattern Recognition
Call for Participation
Group Theory, Invariance and Symmetry in Vision
www.bmva.org/meetings
One-day BMVA symposium on 21st January 2009
Geometrical methods allow analysis of structures and their configuration
in a containing space. Group Theory provides an alternative perspective
analysis of structures, and their containing spaces in terms of
transformations of them. This transformation perspective leads
to the concept of invariant, intrinsic properties; to the concept of
symmetrical structures; allows a radically different characterization of
the geometry of a space; and allows reasoning and modelling in and of
the transformation space itself.
Group theory is a general tool of mathematical analysis, with
applications in most numerate disciplines but with a special
relationship with perception, as first noted by Helmholtz with his ideas
of constancy and later by Poincare who reasoned that perceptual
systems must be able internally to undo the effect of external
transformations.
New applications of Group Theory to Computational Vision continue to be
published. Some, for example understanding statistical shape variation
as a distribution over the group of diffeomorphisms, arise from the
general usefulness of Group Theory; others, for example invariant image
descriptors, are specific to vision.
This meeting is designed as a forum where work in these areas can be
presented and discussed. If you work on any of the topics below or any
related topic, or if you are interested in applying or learning about
them, do participate in the meeting and possibly present your work, even
if it is not fully developed or complete yet.
symmetries of actual and ideal front-end visual systems
invariant image descriptors: local, multilocal and global
perception, detection & analysis of local & extended image symmetry
measurement and use of facial symmetry
invariant flows on images
invariance in colour vision
symmetry-based description of shapes or images for texture or object
recognition
understanding the agent-environment symmetry
symmetry as an indicator of affordance
methods of harmonic analysis for vision
understanding texture as stochastic symmetry
permutation groups in multi-object tracking
applications of groupoids in vision
local symmetry transforms
analysis of shape variation from the transformation perspective
the relationship between image and scene symmetry
Please submit a summary of one A4-sized page (PDF preferred) to Lewis
Griffin ([log in to unmask]) by 12th December 2008.
Dr Dimitrios Makris
Senior Lecturer
Faculty of Computing, Information Systems and Mathematics
Kingston University, London
Tel: +44 20 8517 7082
Email: [log in to unmask]
Web: staffnet.kingston.ac.uk/~ku32195/
This email has been scanned for all viruses by the MessageLabs Email
Security System.
|