Discontinuity in SVD embedding mapping used for watermarks

Kazuo Ohzeki, Yuki Seo, Engyoku Gi

Research output: Contribution to journalArticle

Abstract

It is possible for Singular Value Decomposition (SVD) watermark embedding to have a quasi-one-way functionality. While basic SVD embedding has the drawback of being non-reversible, an advanced SVD method which uses a reversible relation only works one-way. When embedding takes place, the decomposing matrix obtained from the original image differs from the one obtained from the embedded image. This paper demonstrates that there is a distinct difference which shows discontinuity in the Euclidean Norm when evaluated numerically. This evidence points towards the probability that SVD embedding is resilient to inversion attacks.

Original languageEnglish
Pages (from-to)9-17
Number of pages9
JournalInternational Journal of Computer Science and Applications
Volume7
Issue number3
Publication statusPublished - 2010

Fingerprint

Singular value decomposition

Keywords

  • Eigenvalue
  • Inversion attack
  • Mapping
  • Norm
  • One-way function
  • Rank
  • SVD

ASJC Scopus subject areas

  • Computer Science Applications

Cite this

Discontinuity in SVD embedding mapping used for watermarks. / Ohzeki, Kazuo; Seo, Yuki; Gi, Engyoku.

In: International Journal of Computer Science and Applications, Vol. 7, No. 3, 2010, p. 9-17.

Research output: Contribution to journalArticle

Ohzeki, Kazuo ; Seo, Yuki ; Gi, Engyoku. / Discontinuity in SVD embedding mapping used for watermarks. In: International Journal of Computer Science and Applications. 2010 ; Vol. 7, No. 3. pp. 9-17.
@article{1cb23bfe94ea4144b3128650b9aa56f1,
title = "Discontinuity in SVD embedding mapping used for watermarks",
abstract = "It is possible for Singular Value Decomposition (SVD) watermark embedding to have a quasi-one-way functionality. While basic SVD embedding has the drawback of being non-reversible, an advanced SVD method which uses a reversible relation only works one-way. When embedding takes place, the decomposing matrix obtained from the original image differs from the one obtained from the embedded image. This paper demonstrates that there is a distinct difference which shows discontinuity in the Euclidean Norm when evaluated numerically. This evidence points towards the probability that SVD embedding is resilient to inversion attacks.",
keywords = "Eigenvalue, Inversion attack, Mapping, Norm, One-way function, Rank, SVD",
author = "Kazuo Ohzeki and Yuki Seo and Engyoku Gi",
year = "2010",
language = "English",
volume = "7",
pages = "9--17",
journal = "International Journal of Computer Science and Applications",
issn = "0972-9038",
publisher = "Technomathematics Research Foundation",
number = "3",

}

TY - JOUR

T1 - Discontinuity in SVD embedding mapping used for watermarks

AU - Ohzeki, Kazuo

AU - Seo, Yuki

AU - Gi, Engyoku

PY - 2010

Y1 - 2010

N2 - It is possible for Singular Value Decomposition (SVD) watermark embedding to have a quasi-one-way functionality. While basic SVD embedding has the drawback of being non-reversible, an advanced SVD method which uses a reversible relation only works one-way. When embedding takes place, the decomposing matrix obtained from the original image differs from the one obtained from the embedded image. This paper demonstrates that there is a distinct difference which shows discontinuity in the Euclidean Norm when evaluated numerically. This evidence points towards the probability that SVD embedding is resilient to inversion attacks.

AB - It is possible for Singular Value Decomposition (SVD) watermark embedding to have a quasi-one-way functionality. While basic SVD embedding has the drawback of being non-reversible, an advanced SVD method which uses a reversible relation only works one-way. When embedding takes place, the decomposing matrix obtained from the original image differs from the one obtained from the embedded image. This paper demonstrates that there is a distinct difference which shows discontinuity in the Euclidean Norm when evaluated numerically. This evidence points towards the probability that SVD embedding is resilient to inversion attacks.

KW - Eigenvalue

KW - Inversion attack

KW - Mapping

KW - Norm

KW - One-way function

KW - Rank

KW - SVD

UR - http://www.scopus.com/inward/record.url?scp=77957692652&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=77957692652&partnerID=8YFLogxK

M3 - Article

VL - 7

SP - 9

EP - 17

JO - International Journal of Computer Science and Applications

JF - International Journal of Computer Science and Applications

SN - 0972-9038

IS - 3

ER -