An incremental principal component pursuit algorithm via projections onto the ℓ1 ball

Paul Rodriguez; Brendt Wohlberg

doi:10.1109/INTERCON.2017.8079645

An incremental principal component pursuit algorithm via projections onto the ℓ₁ ball

Paul Rodriguez, Brendt Wohlberg

Sección de Electricidad y Electrónica

Los Alamos National Laboratory

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

10 Scopus citations

Abstract

Video background modeling, used to detect moving objects in digital videos, is a ubiquitous pre-processing step in computer vision applications. Principal Component Pursuit (PCP) PCP is among the leading methods for this problem. In this paper we proposed a new convex formulation for PCP, substituting the standard ℓ₁ regularization with a projection onto the ℓ₁-ball. This formulation offers an advantage over the known incremental PCP methods in practical parameter selection and ghosting suppression, while retaining the ability to be implemented in a fully incremental fashion, keeping all the desired properties related to such PCP methods (low memory footprint, adaptation to changes in the background, computational complexity that allows online processing).

Original language	English
Title of host publication	Proceedings of the 2017 IEEE 24th International Congress on Electronics, Electrical Engineering and Computing, INTERCON 2017
Publisher	Institute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)	9781509063628
DOIs	https://doi.org/10.1109/INTERCON.2017.8079645
State	Published - 20 Oct 2017
Event	24th IEEE International Congress on Electronics, Electrical Engineering and Computing, INTERCON 2017 - Cusco, Peru Duration: 15 Aug 2017 → 18 Aug 2017

Publication series

Name	Proceedings of the 2017 IEEE 24th International Congress on Electronics, Electrical Engineering and Computing, INTERCON 2017

Conference

Conference	24th IEEE International Congress on Electronics, Electrical Engineering and Computing, INTERCON 2017
Country/Territory	Peru
City	Cusco
Period	15/08/17 → 18/08/17

Keywords

Incremental Principal Component Pursuit
Video Background Modeling
ℓ Ball projection

Access to Document

10.1109/INTERCON.2017.8079645

Cite this

Rodriguez, P., & Wohlberg, B. (2017). An incremental principal component pursuit algorithm via projections onto the ℓ₁ ball. In Proceedings of the 2017 IEEE 24th International Congress on Electronics, Electrical Engineering and Computing, INTERCON 2017 Article 8079645 (Proceedings of the 2017 IEEE 24th International Congress on Electronics, Electrical Engineering and Computing, INTERCON 2017). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/INTERCON.2017.8079645

Rodriguez, Paul ; Wohlberg, Brendt. / An incremental principal component pursuit algorithm via projections onto the ℓ₁ ball. Proceedings of the 2017 IEEE 24th International Congress on Electronics, Electrical Engineering and Computing, INTERCON 2017. Institute of Electrical and Electronics Engineers Inc., 2017. (Proceedings of the 2017 IEEE 24th International Congress on Electronics, Electrical Engineering and Computing, INTERCON 2017).

@inproceedings{abbcbe90bf3249e8a5b7c28fa7fc5047,

title = "An incremental principal component pursuit algorithm via projections onto the ℓ1 ball",

abstract = "Video background modeling, used to detect moving objects in digital videos, is a ubiquitous pre-processing step in computer vision applications. Principal Component Pursuit (PCP) PCP is among the leading methods for this problem. In this paper we proposed a new convex formulation for PCP, substituting the standard ℓ1 regularization with a projection onto the ℓ1-ball. This formulation offers an advantage over the known incremental PCP methods in practical parameter selection and ghosting suppression, while retaining the ability to be implemented in a fully incremental fashion, keeping all the desired properties related to such PCP methods (low memory footprint, adaptation to changes in the background, computational complexity that allows online processing).",

keywords = "Incremental Principal Component Pursuit, Video Background Modeling, ℓ Ball projection",

author = "Paul Rodriguez and Brendt Wohlberg",

note = "Publisher Copyright: {\textcopyright} 2017 IEEE.; 24th IEEE International Congress on Electronics, Electrical Engineering and Computing, INTERCON 2017 ; Conference date: 15-08-2017 Through 18-08-2017",

year = "2017",

month = oct,

day = "20",

doi = "10.1109/INTERCON.2017.8079645",

language = "English",

series = "Proceedings of the 2017 IEEE 24th International Congress on Electronics, Electrical Engineering and Computing, INTERCON 2017",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

booktitle = "Proceedings of the 2017 IEEE 24th International Congress on Electronics, Electrical Engineering and Computing, INTERCON 2017",

}

Rodriguez, P & Wohlberg, B 2017, An incremental principal component pursuit algorithm via projections onto the ℓ₁ ball. in Proceedings of the 2017 IEEE 24th International Congress on Electronics, Electrical Engineering and Computing, INTERCON 2017., 8079645, Proceedings of the 2017 IEEE 24th International Congress on Electronics, Electrical Engineering and Computing, INTERCON 2017, Institute of Electrical and Electronics Engineers Inc., 24th IEEE International Congress on Electronics, Electrical Engineering and Computing, INTERCON 2017, Cusco, Peru, 15/08/17. https://doi.org/10.1109/INTERCON.2017.8079645

An incremental principal component pursuit algorithm via projections onto the ℓ₁ ball. / Rodriguez, Paul; Wohlberg, Brendt.
Proceedings of the 2017 IEEE 24th International Congress on Electronics, Electrical Engineering and Computing, INTERCON 2017. Institute of Electrical and Electronics Engineers Inc., 2017. 8079645 (Proceedings of the 2017 IEEE 24th International Congress on Electronics, Electrical Engineering and Computing, INTERCON 2017).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - An incremental principal component pursuit algorithm via projections onto the ℓ1 ball

AU - Rodriguez, Paul

AU - Wohlberg, Brendt

PY - 2017/10/20

Y1 - 2017/10/20

N2 - Video background modeling, used to detect moving objects in digital videos, is a ubiquitous pre-processing step in computer vision applications. Principal Component Pursuit (PCP) PCP is among the leading methods for this problem. In this paper we proposed a new convex formulation for PCP, substituting the standard ℓ1 regularization with a projection onto the ℓ1-ball. This formulation offers an advantage over the known incremental PCP methods in practical parameter selection and ghosting suppression, while retaining the ability to be implemented in a fully incremental fashion, keeping all the desired properties related to such PCP methods (low memory footprint, adaptation to changes in the background, computational complexity that allows online processing).

AB - Video background modeling, used to detect moving objects in digital videos, is a ubiquitous pre-processing step in computer vision applications. Principal Component Pursuit (PCP) PCP is among the leading methods for this problem. In this paper we proposed a new convex formulation for PCP, substituting the standard ℓ1 regularization with a projection onto the ℓ1-ball. This formulation offers an advantage over the known incremental PCP methods in practical parameter selection and ghosting suppression, while retaining the ability to be implemented in a fully incremental fashion, keeping all the desired properties related to such PCP methods (low memory footprint, adaptation to changes in the background, computational complexity that allows online processing).

KW - Incremental Principal Component Pursuit

KW - Video Background Modeling

KW - ℓ Ball projection

UR - http://www.scopus.com/inward/record.url?scp=85039984942&partnerID=8YFLogxK

U2 - 10.1109/INTERCON.2017.8079645

DO - 10.1109/INTERCON.2017.8079645

M3 - Conference contribution

AN - SCOPUS:85039984942

T3 - Proceedings of the 2017 IEEE 24th International Congress on Electronics, Electrical Engineering and Computing, INTERCON 2017

BT - Proceedings of the 2017 IEEE 24th International Congress on Electronics, Electrical Engineering and Computing, INTERCON 2017

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 24th IEEE International Congress on Electronics, Electrical Engineering and Computing, INTERCON 2017

Y2 - 15 August 2017 through 18 August 2017

ER -

Rodriguez P, Wohlberg B. An incremental principal component pursuit algorithm via projections onto the ℓ₁ ball. In Proceedings of the 2017 IEEE 24th International Congress on Electronics, Electrical Engineering and Computing, INTERCON 2017. Institute of Electrical and Electronics Engineers Inc. 2017. 8079645. (Proceedings of the 2017 IEEE 24th International Congress on Electronics, Electrical Engineering and Computing, INTERCON 2017). doi: 10.1109/INTERCON.2017.8079645