Home

About

Advanced Search

Browse by Discipline

Scientific Societies

E-print Alerts

Add E-prints

E-print Network
FAQHELPSITE MAPCONTACT US


  Advanced Search  

 
Generative Models for Labeling Multi-Object Configurations in Images
 

Summary: Generative Models for Labeling Multi-Object
Configurations in Images
Yali Amit1
and Alain Trouv´e2
1
Department of Statistics, University of Chicago, Chicago, IL 60637
amit@marx.uchicago.edu
2
CMLA, ENS-Cachan,
61 Ave du President Wilson, 94235 Cachan cedex, France
trouve@cmla.ens-cachan.fr
Abstract. We propose a generative approach to the problem of label-
ing images containing configurations of objects from multiple classes.
The main building blocks are dense statistical models for individual ob-
jects. The models assume conditional independence of binary oriented
edge variables conditional on a hidden instantiation parameter, which
also determines an object support. These models are then be composed
to form models for object configurations with various interactions includ-
ing occlusion. Choosing the optimal configuration is entirely likelihood
based and no decision boundaries need to be pre-learned. Training in-

  

Source: Amit, Yali - Departments of Computer Science & Statistics, University of Chicago
Edelman, Shimon - Departments of Computer Science & Psychology, Cornell University

 

Collections: Biology and Medicine; Computer Technologies and Information Sciences