SCIS & ISIS
SCIS & ISIS 2010
Session ID : TH-F3-1
Conference information
Probabilistic Object Learning through Attention-based Organized Perception
*Masayasu Atsumi
Author information
CONFERENCE PROCEEDINGS FREE ACCESS

Details
Abstract

This paper proposes a model of probabilistic object learning in conjunction with attention-based organized perception. This model consists of the following two sub-models: the former is a model of attention-based perceptual organization of segments and the latter is a model of learning object composition of categories based on a bag of features representation of segments. In attention-based perceptual organization, concurrent figure-ground segmentation is performed on dynamically-formed Markov random fields around salient preattentive points and co-occurring segments of objects and their context are grouped in the neighborhood of selective attended segments. In probabilistic learning of categorical object composition, multi-class classifiers are learned based on intra-categorical probabilistic latent component analysis with variable number of classes and inter-categorical typicality analysis. Through experiments by using images of plural categories in an image database, it is shown that the model learns a probabilistic structure of intra-categorical composition of objects and context and inter-categorical difference.

Content from these authors
© 2010 Japan Society for Fuzzy Theory and Intelligent Informatics
Previous article Next article
feedback
Top