[comp.ai.neural-nets] Intell. Systems for Signal and Image Understanding

unige@cernvax.cern.ch (nicolas mayencourt) (01/31/91)

 
 
           PRELIMINARY ANNOUNCEMENT AND CALL FOR PAPERS
 
  COURSE ON INTELLIGENT SYSTEMS FOR SIGNAL AND IMAGE UNDERSTANDING
 
 
PLACE AND DATE:
 
Udine (Italy), International Center for Mechanical Sciences (CISM),
October 1-5, 1991;
 
SPONSORS:
 
EURASIP (European Association for Signal Processing); UNESCO;
University of Udine;
 
MOTIVATIONS AND SCOPE:
 
By 'Intelligent System' is meant a collection of modules (hardware,
software) aimed at emulating the ways humans adopt to solve problems.
It is now widely recognised that problems involving perceptive
capabilities (visual, auditory, tactile,..), and their integration
in activities such as scene understanding, speech recognition,
signal and data interpretation, senso-motorial coordination, are
particularly difficult to emulate.
The course introduces the basic techniques that are
currently used in the design of intelligent systems for signal and image
understanding: terminology and concepts will be presented
concerning the knowledge-based approach, as well as the
connectionist paradigms.
Attention will be devoted to various aspects of 'integration' (numerical
with symbolic representations and processing steps, data fusion from sensors)
in suitable architectural and programming schemes; case studies will be
discussed, concerning applications in signal and image domains.
 
The sessions of the school will contain tutorial lectures and oral
presentations.
Practical software demonstrations will also take place.
 
WHO SHOULD ATTEND:
 
University researchers and graduate students in electrical engineering,
computer science and related; practising engineers and technical managers
involved in signal and image analysis in industrial domains.
Experience in the use of computers is desirable.
Participation is limited.
 
MAIN TOPICS:
 
1.- Intelligent system architectures, with emphasis to signal
understanding and image interpretation: blackboard systems, distributed
(multi-agent) systems;
 
2.- Non-numerical programming techniques: functional, logical,
object-oriented; integration of numerical and symbolic programming
modules;
 
3.- Multi-sensor data fusion techniques (statistical; inferential);
 
4.- Visual modelling techniques; matching and reasoning in visual
systems;
 
5.- Parallel distributed systems; neural computing techniques;
an overview on signal- and image-oriented applications;
 
6.- Case studies.
 
A detailed timetable of the lectures will be sent to the participants.
 
INVITED LECTURERS:
 
- Y. Demazeau, LIFIA - CNRS, Grenoble (F);
 
- C. Liedtke, University of Hannover (D);
 
- C. Tasso, University of Udine (I);
 
- E. Trucco, University of Genoa (I) and Edinburgh (UK);
 
- D. Willshaw, University of Edinburgh (UK);
 
PROCEEDINGS:
 
Will be published on a special issue of SIGNAL PROCESSING (Elsevier);
papers will be refereed.
 
Those participants who wish to submit a paper, are requested to send four
copies of their contribution, in its final form, to the Coordinator of the
Course by June 1, 1991.
 
COORDINATOR:
 
- prof. Vito ROBERTO;
 
Dipartimento di Inform
atica,
Via Zanon, 6 UDINE (Italy)
 
Telefax +39 432 510755; Telephone: +39 432 297169;
 
E-mail: roberto@uduniv.cineca.it
 
FURTHER INFORMATION: contact the Coordinator by e-mail or fax.