Shape and Motion


SAM - EC-ISR93003

Keywords: robotics, object recognition, active vision, motion estimation


Start Date: 1 November 93 / Duration: 42 months

[ participants / contact]


Objectivies and Approach

The objective of this cooperation is to investigate robust techniques for continuous computation of shape primitives that facilitate interpretation of visual, static as well as dynamic scenes. The investigations will include both theoretical analyses and test implementations of these techniques in integrated and continuously operating systems.

The European teams have developed continuously operating systems in ESPRIT BRA VAP-II. The Israeli teams have a high level of expertise in methods. This collaboration provides opportunities to refine methods and systems on the basis of the joint experience. This includes testing and refining methods in system contexts, and achieving more robust system performance through integration of improved techniques.

European links: ESPRIT BRA VAP-II (7108), Network RETINA under the Human Capital Mobility Programme, EC-US003.

Progress and Results

Intensive investigations are being carried out and progress is being made towards the development of a framework that unifies model-based and appearance-based approaches by exploiting the best of them. Regarding view-planning, optimal viewpoints have been defined from apriori geometrical models, and strategies have been developed for controlling active cameras to noise-insensitive views without knowing models of the observed object. The control framework "Active Vision Shell" for intelligent Perception-Action processes is a direct result of the collaboration.

The diversity of the scientific approaches of the teams involved has also inspired the development of a range of new "hybrid" methods for indexing and accumulation of evidence, foveated vision, and selection of spatial scales for various purposes of analysis.

Information Dissemination Activities and/or Exploitation

Results have been presented at international conferences and annual workshops jointly organised by SAM and the working group EC-US003.

The integration of alternative and improved methods into the active vision systems will result in the development of a range of new and robust techniques that will be relevant for industrial applications for which this robustness is crucial.


Coordinator

Aalborg University
Laboratory of Image Analysis
Fredrik Bajers Vej 7, Bldg. D-1
DK - 9220 AALBORG, DK

EU Partners

Aalborg University, DK
Institut Polytechnique de Grenoble, F
Università degli Studi di Genova, I
Linköping University, S
Royal Institute of Technology, S
University of Surrey, UK

Non-EU Partners

The Hebrew University of Jerusalem, ISR
Weizmann Institute of Science, ISR
The Technion, ISR
Tel Aviv University, ISR

CONTACT POINT

Prof. Erik Granum
Tel: +45 98 15 11 33
Fax: +44 98 15 40 08
E-mail: eg@vision.auc.dk


INCO synopses home page INCO cooperation index INCO keyword index
INCO acronym index INCO number index INCO Projects index
All
synopses home page all keywords index all acronyms index all numbers index

SAM - EC-ISR93003, May 1997


please address enquiries to the ESPRIT Information Desk

html version of synopsis by Nick Cook