Spotlight the Negatives: A Generalized Discriminative Latent Model

Hossein Azizpour, Mostafa Arefiyan, Sobhan Naderi Parizi and Stefan Carlsson

Abstract

Discriminative latent variable models (LVM) are frequently applied to various visual recognition tasks. In these systems the latent (hidden) variables provide a formalism for modeling structured variation of visual features. Conventionally, latent variables are defined on the variation of the foreground (positive) class. In this work we augment LVMs to include \textit{negative} latent variables corresponding to the background class. We formalize the scoring function of such a generalized LVM (GLVM). Then we discuss a framework for learning a model based on the GLVM scoring function. We theoretically showcase how some of the current visual recognition methods can benefit from this generalization. Finally, we experiment on a generalized form of Deformable Part Models with negative latent variables and show significant improvements on two different detection tasks.

Session

Poster 1

Files

Extended Abstract (PDF, 349K)
Paper (PDF, 1285K)

DOI

10.5244/C.29.18
https://dx.doi.org/10.5244/C.29.18

Citation

Bibtex

@inproceedings{BMVC2015_18,
	title={Spotlight the Negatives: A Generalized Discriminative Latent Model},
	author={Hossein Azizpour and Mostafa Arefiyan and Sobhan Naderi Parizi and Stefan Carlsson},
	year={2015},
	month={September},
	pages={18.1-18.11},
	articleno={18},
	numpages={11},
	booktitle={Proceedings of the British Machine Vision Conference (BMVC)},
	publisher={BMVA Press},
	editor={Xianghua Xie, Mark W. Jones, and Gary K. L. Tam},
	doi={10.5244/C.29.18},
	isbn={1-901725-53-7},
	url={https://dx.doi.org/10.5244/C.29.18}
}