Grant: Discovery Grants Program - Individual
Carleton University
$19,000.00
Visuospatial scene generation
EF 2011-12
Résumé
I propose to create a system that finds statistical regularities in databases of labeled images, as well as quantitative ranges of qualitative descriptors (e.g. fairly tall men are between 5'10"" and 6'4"" in height) to create a visuospatial scene generation (VSG) model. The model will output a symbolic scene description that includes the following information: 1) what objects are present in the scene, 2) which particular images to use to represent those objects, 3) where those objects are, both in terms of absolute location and relative to other objects, and 4) quantitative magnitudes of input adjectives and prepositions (such as ""large"" or ""above""). More specifically, it will output these things in terms of probability distributions.
This model will be used in several ways: 1) it can be used as a part of a generate-and-test algorithm for computer vision, as other generative models have done. 2) it can be used to suggest to computer vision systems what other objects might be likely to appear in an image, and indeed where they might be in the image. These suggestions will improve the vision system's final classification decisions. 3) the system can be used to create actual images for people to view. In the longer term, tools such as these would provide inspiration to creative illustrators working on open-ended problems: for example, those who create posters, illustrations, and other kinds of graphical layouts. Additionally, it will be a first pass at a model of human visual imagination.
The system will be evaluated with use of object recognition programs, and through user testing.
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
341471-2009
Name nom
Davies, Jim
Department département
Cognitive Science, Institute of
Organization id
23
Institution établissement
Carleton University
Province en
Ontario
Province fr
Ontario
Country en
CANADA
Country fr
CANADA
Fiscal year exercice financier
2011
Competition year année de concours
2009
Award amount
19000
Program id
RGPIN
Program nane 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
331
Committee name en
Computing and Information Sciences - B
Committee name fr
Informatique - B
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
2800
Research subject group en
Artificial intelligence
Research subject group fr
Intelligence artificielle
Research subject en
Artificial intelligence (computer vision, use2603)
Research subject fr
Intelligence artificielle (vision artificielle, utiliser 2603)
Partie
3
Nb partie
5
Application title
Visuospatial scene generation
Keyword
NSERC
Application summary
I propose to create a system that finds statistical regularities in databases of labeled images, as well as quantitative ranges of qualitative descriptors (e.g. fairly tall men are between 5'10"" and 6'4"" in height) to create a visuospatial scene generation (VSG) model. The model will output a symbolic scene description that includes the following information: 1) what objects are present in the scene, 2) which particular images to use to represent those objects, 3) where those objects are, both in terms of absolute location and relative to other objects, and 4) quantitative magnitudes of input adjectives and prepositions (such as ""large"" or ""above""). More specifically, it will output these things in terms of probability distributions.
This model will be used in several ways: 1) it can be used as a part of a generate-and-test algorithm for computer vision, as other generative models have done. 2) it can be used to suggest to computer vision systems what other objects might be likely to appear in an image, and indeed where they might be in the image. These suggestions will improve the vision system's final classification decisions. 3) the system can be used to create actual images for people to view. In the longer term, tools such as these would provide inspiration to creative illustrators working on open-ended problems: for example, those who create posters, illustrations, and other kinds of graphical layouts. Additionally, it will be a first pass at a model of human visual imagination.
The system will be evaluated with use of object recognition programs, and through user testing.