A Fast Weighted SVT Algorithm

Aritra Dutta; Xin Li

doi:10.1109/icsai.2018.8599289

A Fast Weighted SVT Algorithm

Aritra Dutta, Xin Li

Computer, Electrical and Mathematical Sciences and Engineering

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

3 Scopus citations

Abstract

Singular value thresholding (SVT) plays an important role in the well-known robust principal component analysis (RPCA) algorithms which have many applications in machine learning, pattern recognition, and computer vision. There are many versions of generalized SVT proposed by researchers to achieve improvement in speed or performance. In this paper, we propose a fast algorithm to solve aweighted singular value thresholding (WSVT) problem as formulated in [1], which uses a combination of the nuclear norm and a weighted Frobenius norm and has shown to be comparable with RPCA method in some real world applications.

Original language	English (US)
Title of host publication	2018 5th International Conference on Systems and Informatics (ICSAI)
Publisher	Institute of Electrical and Electronics Engineers (IEEE)
Pages	1022-1026
Number of pages	5
ISBN (Print)	9781728101200
DOIs	https://doi.org/10.1109/icsai.2018.8599289
State	Published - Jan 3 2019

Access to Document

10.1109/icsai.2018.8599289

Cite this

@inproceedings{46f66267c1e94c07a58842cf77d0d507,

title = "A Fast Weighted SVT Algorithm",

abstract = "Singular value thresholding (SVT) plays an important role in the well-known robust principal component analysis (RPCA) algorithms which have many applications in machine learning, pattern recognition, and computer vision. There are many versions of generalized SVT proposed by researchers to achieve improvement in speed or performance. In this paper, we propose a fast algorithm to solve aweighted singular value thresholding (WSVT) problem as formulated in [1], which uses a combination of the nuclear norm and a weighted Frobenius norm and has shown to be comparable with RPCA method in some real world applications.",

author = "Aritra Dutta and Xin Li",

note = "KAUST Repository Item: Exported on 2020-10-01",

year = "2019",

month = jan,

day = "3",

doi = "10.1109/icsai.2018.8599289",

language = "English (US)",

isbn = "9781728101200",

pages = "1022--1026",

booktitle = "2018 5th International Conference on Systems and Informatics (ICSAI)",

publisher = "Institute of Electrical and Electronics Engineers (IEEE)",

}

TY - GEN

T1 - A Fast Weighted SVT Algorithm

AU - Dutta, Aritra

AU - Li, Xin

N1 - KAUST Repository Item: Exported on 2020-10-01

PY - 2019/1/3

Y1 - 2019/1/3

N2 - Singular value thresholding (SVT) plays an important role in the well-known robust principal component analysis (RPCA) algorithms which have many applications in machine learning, pattern recognition, and computer vision. There are many versions of generalized SVT proposed by researchers to achieve improvement in speed or performance. In this paper, we propose a fast algorithm to solve aweighted singular value thresholding (WSVT) problem as formulated in [1], which uses a combination of the nuclear norm and a weighted Frobenius norm and has shown to be comparable with RPCA method in some real world applications.

AB - Singular value thresholding (SVT) plays an important role in the well-known robust principal component analysis (RPCA) algorithms which have many applications in machine learning, pattern recognition, and computer vision. There are many versions of generalized SVT proposed by researchers to achieve improvement in speed or performance. In this paper, we propose a fast algorithm to solve aweighted singular value thresholding (WSVT) problem as formulated in [1], which uses a combination of the nuclear norm and a weighted Frobenius norm and has shown to be comparable with RPCA method in some real world applications.

UR - http://hdl.handle.net/10754/630940

UR - https://ieeexplore.ieee.org/document/8599289

UR - http://www.scopus.com/inward/record.url?scp=85061510605&partnerID=8YFLogxK

U2 - 10.1109/icsai.2018.8599289

DO - 10.1109/icsai.2018.8599289

M3 - Conference contribution

SN - 9781728101200

SP - 1022

EP - 1026

BT - 2018 5th International Conference on Systems and Informatics (ICSAI)

PB - Institute of Electrical and Electronics Engineers (IEEE)

ER -

A Fast Weighted SVT Algorithm

Abstract

Access to Document

Other files and links

Fingerprint

Cite this