Electrical Engineering
      and Computer Sciences

Electrical Engineering and Computer Sciences

COLLEGE OF ENGINEERING

UC Berkeley

Don't Look Back: Post-hoc Category Detection via Sparse Reconstruction

Hyun Oh Song, Mario Fritz, Tim Althoff and Trevor Darrell

EECS Department
University of California, Berkeley
Technical Report No. UCB/EECS-2012-16
January 24, 2012

http://www.eecs.berkeley.edu/Pubs/TechRpts/2012/EECS-2012-16.pdf

We consider optimal representations for representing prototypical categories in the latent deformable part model framework, with a specific emphasis on category-level retrieval tasks defined "on the fly'' for a large corpus. In this setting, it is impractical to perform an exhaustive search with a full model; we investigate methods which approximately reconstruct the score function of a novel category from a set of precomputed responses. We propose a novel sparse reconstruction method where part classifiers are decomposed via a shared dictionary of part filters; in turn, our method can efficiently reconstruct approximate part responses on large image corpora using a sparse matrix-vector product based on pre-computed filter responses instead of exhaustive convolutions. We compare our method to baseline schemes using SVD-based or nearest-category approximation and show our method is more effective at detecting novel categories. We additionally demonstrate results towards an end-to-end system for activity detection which trains a protoype category concept model from one dataset (PASCAL), learns post-hoc categories on the fly based on training data from a second dataset where labeled data are available (ImageNet), and sucessfully detects instances in test data from a third dataset (TRECVID MED) via reconstruction with the precomputed prototype models.

Author Comments: Contact Trevor Darrell trevor@eecs for information.


BibTeX citation:

@techreport{Song:EECS-2012-16,
    Author = {Song, Hyun Oh and Fritz, Mario and Althoff, Tim and Darrell, Trevor},
    Title = {Don't Look Back: Post-hoc Category Detection via Sparse Reconstruction},
    Institution = {EECS Department, University of California, Berkeley},
    Year = {2012},
    Month = {Jan},
    URL = {http://www.eecs.berkeley.edu/Pubs/TechRpts/2012/EECS-2012-16.html},
    Number = {UCB/EECS-2012-16},
    Note = {Contact Trevor Darrell trevor@eecs for information.},
    Abstract = {We consider optimal representations for representing prototypical categories in the latent deformable part model framework, with a specific emphasis on category-level retrieval tasks defined "on the fly'' for a large corpus. In this setting, it is impractical to perform an exhaustive search with a full model; we investigate methods which approximately reconstruct the score function of a novel category from a set of precomputed responses.  We propose a novel sparse reconstruction method where part classifiers are decomposed via a shared dictionary of part filters; in turn, our method can efficiently reconstruct approximate part responses on large image corpora using a sparse matrix-vector product based on pre-computed filter responses instead of exhaustive convolutions.  We compare our method to baseline schemes using SVD-based or nearest-category approximation and show our method is more effective at detecting novel categories.  We additionally demonstrate results towards an end-to-end system for activity detection which trains a protoype category concept model from one dataset (PASCAL), learns post-hoc categories on the fly based on training data from a second dataset where labeled data are available (ImageNet), and sucessfully detects instances in test data from a third dataset (TRECVID MED) via reconstruction with the precomputed prototype models.}
}

EndNote citation:

%0 Report
%A Song, Hyun Oh
%A Fritz, Mario
%A Althoff, Tim
%A Darrell, Trevor
%T Don't Look Back: Post-hoc Category Detection via Sparse Reconstruction
%I EECS Department, University of California, Berkeley
%D 2012
%8 January 24
%@ UCB/EECS-2012-16
%U http://www.eecs.berkeley.edu/Pubs/TechRpts/2012/EECS-2012-16.html
%F Song:EECS-2012-16