# Online and Batch Supervised Background Estimation via L1 Regression

Handle URI:
http://hdl.handle.net/10754/626534
Title:
Online and Batch Supervised Background Estimation via L1 Regression
Authors:
Dutta, Aritra; Richtarik, Peter
Abstract:
We propose a surprisingly simple model for supervised video background estimation. Our model is based on $\ell_1$ regression. As existing methods for $\ell_1$ regression do not scale to high-resolution videos, we propose several simple and scalable methods for solving the problem, including iteratively reweighted least squares, a homotopy method, and stochastic gradient descent. We show through extensive experiments that our model and methods match or outperform the state-of-the-art online and batch methods in virtually all quantitative and qualitative measures.
KAUST Department:
Computer, Electrical and Mathematical Sciences and Engineering (CEMSE) Division; Computer Science Program
Publisher:
arXiv
Issue Date:
23-Nov-2017
ARXIV:
arXiv:1712.02249
Type:
Preprint
http://arxiv.org/abs/1712.02249v1; http://arxiv.org/pdf/1712.02249v1
Appears in Collections:
Other/General Submission; Computer Science Program; Computer, Electrical and Mathematical Sciences and Engineering (CEMSE) Division

DC FieldValue Language
dc.contributor.authorDutta, Aritraen
dc.contributor.authorRichtarik, Peteren
dc.date.accessioned2017-12-28T07:32:15Z-
dc.date.available2017-12-28T07:32:15Z-
dc.date.issued2017-11-23en
dc.identifier.urihttp://hdl.handle.net/10754/626534-
dc.description.abstractWe propose a surprisingly simple model for supervised video background estimation. Our model is based on $\ell_1$ regression. As existing methods for $\ell_1$ regression do not scale to high-resolution videos, we propose several simple and scalable methods for solving the problem, including iteratively reweighted least squares, a homotopy method, and stochastic gradient descent. We show through extensive experiments that our model and methods match or outperform the state-of-the-art online and batch methods in virtually all quantitative and qualitative measures.en
dc.publisherarXiven
dc.relation.urlhttp://arxiv.org/abs/1712.02249v1en
dc.relation.urlhttp://arxiv.org/pdf/1712.02249v1en
dc.rightsArchived with thanks to arXiven
dc.titleOnline and Batch Supervised Background Estimation via L1 Regressionen
dc.typePreprinten
dc.contributor.departmentComputer, Electrical and Mathematical Sciences and Engineering (CEMSE) Divisionen
dc.contributor.departmentComputer Science Programen
dc.eprint.versionPre-printen
dc.contributor.institutionMIPTen
dc.contributor.institutionEdinburghen
dc.identifier.arxividarXiv:1712.02249en
kaust.authorDutta, Aritraen
kaust.authorRichtarik, Peteren