1st International ICST Conference on Forensic Applications and Techniques in Telecommunications, Information and Multimedia

Research Article

SVD-Based Watermark with Quasi-One-Way Operation by Reducing a Singular Value Matrix Rank

  • @INPROCEEDINGS{10.4108/e-forensics.2008.2672,
        author={Kazuo Ohzeki and Masaru Sakurai},
        title={SVD-Based Watermark with Quasi-One-Way Operation by Reducing a Singular Value Matrix Rank},
        proceedings={1st International ICST Conference on Forensic Applications and Techniques in Telecommunications, Information and Multimedia},
        publisher={ACM},
        proceedings_a={E-FORENSICS},
        year={2010},
        month={5},
        keywords={watermark SVD quasi-one-way function inversion},
        doi={10.4108/e-forensics.2008.2672}
    }
    
  • Kazuo Ohzeki
    Masaru Sakurai
    Year: 2010
    SVD-Based Watermark with Quasi-One-Way Operation by Reducing a Singular Value Matrix Rank
    E-FORENSICS
    ACM
    DOI: 10.4108/e-forensics.2008.2672
Kazuo Ohzeki1,*, Masaru Sakurai2,*
  • 1: Shibaura Institute of Technology ISE, 3-7-5, Toyosu, Koutou-ku, Tokyo 135-8548 +81-3-5859-8513
  • 2: Nagoya Institute of Technology CSE, Gokiso-cho, Showa-ku, Nagoya 466-8555
*Contact email: ohzeki@sic.shibaura-it.ac.jp, sakurai.masaru@nitech.ac.jp

Abstract

A new Singular Value Decomposition (SVD)-based watermarking method was developed in this study. Quasione- way operation is improved by reducing the rank of the singular value matrix. All diagonal elements of the singular value matrix are positive; all other off-diagonal elements are zeros. To add an element to the off-diagonal area is fundamentally a quasi-one-way operation because no value cannot be subtracted from zero. In adding elements to the off-diagonal area, a pair of linearly dependent vectors is produced. This modification reduces the rank of the singular value matrix. It is computationally difficult to find diminished singular values from the reduced rank matrix.