CVSL Logo
FrancaisHome
AboutPeopleResearchPublicationsEventsProfile
About
COGNOIS

 

 

 

REPARTI

MIVIM

COGNOIS:

Communication and Observation toward a Generic Natural Ontogeny for Intelligent Systems

The ultimate goal of COGNOIS is to develop a generic computing architecture leading to a dependable intelligent machine endowed with the required observation and communication abilities and readily amendable to various real-world applications. A major challenge in developing dependable intelligent machines exploiting extensive behavioural modeling hinges upon a fully integrated exploitation of a number of sensing modalities - optical, acoustical, mechanical, etc. - and the validation of the respective (visual, auditory, haptic, etc.) knowledge, using a variety of human-machine communication channels. Other challenges involve data selection and segmentation and knowledge representation and summarizing.

COGNOIS builds upon the “Actors-Properties-Interactions-Architecture” (APIA) which is under development in VERTEX. COGNOIS seeks ways to model general behaviours (ex. how does a human expert operator perform a delicate, dextrous task). This project focuses on a specific multimodal observation framework combining computer vision and speech recognition and integrates derived behaviours within virtual environments using advanced human-machine interfaces. The operating scenario of the COGNOIS architecture involves the following three steps:

  1. Knowledge acquisition and validation of significant entities, building on existing approaches and learning to deal with their deficiencies;

  2. Knowledge acquisition and validation of significant behaviours, stemming from known entities in the environment;

  3. Knowledge use, i.e. documentation of the observations, with the goal of developing a working prototype architecture which is first applied to situations involving a small number of simple entities and is later refined as the complexity of the observed situations is gradually increased.


More...

Project title    Investigator(s)    Level
   Incremental description of deformable objects in video sequences using computer vision    Stéphane Drouin (Student)
Marc Parizeau (Supervisor)
Patrick Hébert (Co-supervisor)
   Ph.D.
(1/2003 - 12/2007)
   Detecting humans in a cluttered scene    Jérôme Vignola (Student)
Robert Bergevin (Supervisor)
   Masters
(9/2001 - 12/2003)
   Region matching of people images taken at different times and locations    Michel Lantagne (Student)
Robert Bergevin (Supervisor)
Marc Parizeau (Co-supervisor)
   Masters
(9/2001 - 9/2003)
   Person Identification System with Computer Vision    Alexandre Lemieux (Student)
Marc Parizeau (Supervisor)
   Masters
(9/2000 - 6/2003)
   Description of a moving person observed by a multi-camera system    Stéphane Drouin (Student)
Marc Parizeau (Supervisor)
Patrick Hébert (Co-supervisor)
   Masters
(5/2001 - 12/2002)

Publications

     
   
   

©2002-. Computer Vision and Systems Laboratory. All rights reserved