Paper Info
Title
Contextual classification of image patches with latent aspect models
Abstract
We present a novel approach for contextual classification of image patches in complex visual scenes, based on the use of histograms of quantized features and probabilistic aspect models. Our approach uses context in two ways: (1) by using the fact that specific learned aspects correlate with the semantic classes, which resolves some cases of visual polysemy often present in patch-based representations, and (2) by formalizing the notion that scene context is image-specific--what an individual patch represents depends on what the rest of the patches in the same image are. We demonstrate the validity of our approach on aman-made versus natural patch classification problem. Experiments on an image collection of complex scenes show that the proposed approach improves region discrimination, producing satisfactory results and outperforming two noncontextual methods. Furthermore, we also show that co-occurrence and traditional (Markov random field) spatial contextual information can be conveniently integrated for further improved patch classification.
Year
DOI
Venue
2009
10.1155/2009/602920
EURASIP J. Image and Video Processing
Keywords
Field
DocType
improved patch classification,natural patch classification problem,novel approach,latent aspect model,complex scene,complex visual scene,individual patch,image patch,contextual classification,image collection,object recognition,segmentation
Scale-invariant feature transform,Latent Dirichlet allocation,Computer science,Markov random field,Artificial intelligence,Probabilistic logic,Contextual image classification,Computer vision,Pattern recognition,Segmentation,Mixture model,Machine learning,Cognitive neuroscience of visual object recognition
Journal
Volume
Issue
ISSN
2009,
1
1687-5281
Citations 
PageRank 
References 
3
0.54
33
Authors
4
Name
Order
Citations
PageRank
Florent Monay159331.43
Pedro Quelhas226121.51
Jean-marc Odobez31641110.52
Daniel Gatica-Perez44182276.74