DMCA
Efficient learning of relational object class models (2005)
| Venue: | In ICCV |
| Citations: | 26 - 7 self |
BibTeX
@INPROCEEDINGS{Bar-hillel05efficientlearning,
author = {Aharon Bar-hillel and Daphna Weinshall},
title = {Efficient learning of relational object class models},
booktitle = {In ICCV},
year = {2005},
pages = {1762--1769}
}
Share
 |
 |
 |
 |
||
OpenURL
Abstract
We present an efficient method for learning part-based object class models. The models include part appearance, as well as location and scale relations between parts. The object class is generatively modeled using a simple Bayesian network with a central hidden node containing location and scale information, and nodes describing object parts. The model’s parameters, however, are optimized to reduce a loss function of the training error, as in discriminative methods. We show how boosting techniques can be extended to optimize the relational model proposed, with complexity linear in the number of parts and the number of features per image. This efficiency allows our method to learn relational models with many parts and features. The method has an advantage over purely generative and purely discriminative approaches, since the former are limited to a small number of parts and features, while the latter neglect geometrical relations between parts. Experimental results are described, using some bench-mark data sets and three sets of newly collected data, showing the relative merits of our method in recognition and localization tasks. 1
Keyphrases
efficient learning relational object class model relational model simple bayesian network latter neglect geometrical relation loss function discriminative approach central hidden node discriminative method training error part-based object class model object part part appearance efficient method model parameter scale relation localization task scale information many part small number object class bench-mark data set relative merit experimental result
