Canada Spends Logo

Explorez notre analyse du budget 2026

Grant: Discovery Grants Program - Individual

University of Ottawa

$25,000.00

Multi-view computer vision

EF 2007-08

Résumé
The recent convergence of the video, vision and graphics research fields has opened the door to a broad range of new exciting applications. The proposed research would like to contribute to this effervescence by proposing and devising new approaches based on multi-camera vision systems. The research work will be subdivided into two projects, each addressing different complementary aspects of the multi-camera vision problematic: real-time scene augmentation and continous scene modeling and monitoring. In augmented reality, we will pursue our research on real-time scene augmentation based on the collaboration between a moving camera and a set of fixed reference cameras. Our tensor-based real-time system for scene augmentation has been shown to work effectively with one moving camera and two fixed cameras. The next step is now to extend it to the case of multiple views covering a larger scene. Another important extension of our work will consist in replacing the single moving camera by a stereo pair. This camera pair will allow us to deal with occlusions, i.e. to have the inserted virtual object occluded when it passes behind a real object. The second project concerns the case where the 3D structure of an observed scene is known a priori. In particular, we want to study how a 3D model of a complex scene can be kept consistent with the observations as this scene evolves in time (e.g. some objects being moved or new objects added). The scene is observed by several cameras. The model will describe the structure of the observed elements but not their appearance; visual appearance will have to be inferred from the observations. When an object is moved inside the scene, the 3D model will be automatically updated by finding the rigid displacement that best explain the new visual information. When a new object will be added to the scene, the different available views will be used to localize it.
Programme
Discovery Grants Program - Individual
Bénéficiaire
Emplacement
Ontario, Canada
Type de bénéficiaire
Academia
Type
Grant
Source
NSERC's Awards Data

Tel que publié

Application id
184103-2006
Name nom
Laganière, Robert
Department département
Engineering, Faculty of
Organization id
28
Institution établissement
University of Ottawa
Province en
Ontario
Province fr
Ontario
Country en
CANADA
Country fr
CANADA
Fiscal year exercice financier
2007
Competition year année de concours
2006
Award amount
25000
Program id
RGPIN
Program name en
Discovery Grants Program - Individual
Program name fr
Programme de subventions à la découverte - individuelles
Group en
Discovery grants programs
Group fr
Programmes de subventions à la découverte
Committee code
335
Committee name en
Electromag. & Electr. Systems (H)
Committee name fr
Électromagn. et syst. électr. (H)
Area of application code
802
Area of application group en
Information and communication services
Area of application group fr
Services d'information et de télécommunications
Area of application en
Information systems and technology
Area of application fr
Systèmes d'information et technologie de l'information
Research subject code
2603
Research subject group en
Robotics
Research subject group fr
Robotique
Research subject en
Computer vision
Research subject fr
Vision artificielle
Partie
2
Nb partie
5
Application title
Multi-view computer vision
Keyword
NA
Application summary
The recent convergence of the video, vision and graphics research fields has opened the door to a broad range of new exciting applications. The proposed research would like to contribute to this effervescence by proposing and devising new approaches based on multi-camera vision systems. The research work will be subdivided into two projects, each addressing different complementary aspects of the multi-camera vision problematic: real-time scene augmentation and continous scene modeling and monitoring. In augmented reality, we will pursue our research on real-time scene augmentation based on the collaboration between a moving camera and a set of fixed reference cameras. Our tensor-based real-time system for scene augmentation has been shown to work effectively with one moving camera and two fixed cameras. The next step is now to extend it to the case of multiple views covering a larger scene. Another important extension of our work will consist in replacing the single moving camera by a stereo pair. This camera pair will allow us to deal with occlusions, i.e. to have the inserted virtual object occluded when it passes behind a real object. The second project concerns the case where the 3D structure of an observed scene is known a priori. In particular, we want to study how a 3D model of a complex scene can be kept consistent with the observations as this scene evolves in time (e.g. some objects being moved or new objects added). The scene is observed by several cameras. The model will describe the structure of the observed elements but not their appearance; visual appearance will have to be inferred from the observations. When an object is moved inside the scene, the 3D model will be automatically updated by finding the rigid displacement that best explain the new visual information. When a new object will be added to the scene, the different available views will be used to localize it.