Ashish Kapoor and Raquel Urtasun and Trevor Darrell

EECS Department, University of California, Berkeley

Technical Report No. UCB/EECS-2009-16

January 27, 2009

http://www2.eecs.berkeley.edu/Pubs/TechRpts/2009/EECS-2009-16.pdf

Recognition of general visual categories requires a diverse set of feature types, but not all are equally relevant to individual categories; efficient recognition arises by learning the potentially sparse features for each class and understanding the relationship between features common to related classes. This paper describes hierarchical discriminative probabilistic techniques for learning visual object category models. Our method recovers a nested set of object categories with chosen kernel combinations for discrimination at each level of the tree. We use a Gaussian Process based framework, with a parameterized sparsity penalty to favor compact classification hierarchies. We exploit structural properties of Gaussian Processes in a multi-class setting to gain computational efficiency and employ evidence maximization to optimally infer kernel weights from training data. Experiments on benchmark datasets show that our hierarchical probabilistic kernel combination scheme offers a benefit in both computational efficiency and performance: we report a significant improvement in accuracy compared to the current best whole-image kernel combination schemes on Caltech 101, as well as a two order-ofmagnitude improvement in efficiency.

---

BibTeX citation:

@techreport{Kapoor:EECS-2009-16,
    Author= {Kapoor, Ashish and Urtasun, Raquel and Darrell, Trevor},
    Title= {Probabilistic Kernel Combination for Hierarchical Object Categorization},
    Year= {2009},
    Month= {Jan},
    Url= {http://www2.eecs.berkeley.edu/Pubs/TechRpts/2009/EECS-2009-16.html},
    Number= {UCB/EECS-2009-16},
    Abstract= {Recognition of general visual categories requires a diverse
set of feature types, but not all are equally relevant to
individual categories; efficient recognition arises by learning
the potentially sparse features for each class and understanding
the relationship between features common to
related classes. This paper describes hierarchical discriminative
probabilistic techniques for learning visual object
category models. Our method recovers a nested set of object
categories with chosen kernel combinations for discrimination
at each level of the tree. We use a Gaussian Process
based framework, with a parameterized sparsity penalty to
favor compact classification hierarchies. We exploit structural
properties of Gaussian Processes in a multi-class setting
to gain computational efficiency and employ evidence
maximization to optimally infer kernel weights from training
data. Experiments on benchmark datasets show that
our hierarchical probabilistic kernel combination scheme
offers a benefit in both computational efficiency and performance:
we report a significant improvement in accuracy
compared to the current best whole-image kernel combination
schemes on Caltech 101, as well as a two order-ofmagnitude
improvement in efficiency.},
}

EndNote citation:

%0 Report
%A Kapoor, Ashish 
%A Urtasun, Raquel 
%A Darrell, Trevor 
%T Probabilistic Kernel Combination for Hierarchical Object Categorization
%I EECS Department, University of California, Berkeley
%D 2009
%8 January 27
%@ UCB/EECS-2009-16
%U http://www2.eecs.berkeley.edu/Pubs/TechRpts/2009/EECS-2009-16.html
%F Kapoor:EECS-2009-16