• DocumentCode
    3494073
  • Title

    A colour image watermarking technique resistant to affine geometric attacks

  • Author

    Owalla, Felix O. ; Mwangi, Elijah

  • Author_Institution
    Dept. of Electr. & Inf. Eng., Univ. of Nairobi, Nairobi, Kenya
  • fYear
    2013
  • fDate
    9-12 Sept. 2013
  • Firstpage
    1
  • Lastpage
    5
  • Abstract
    This paper presents a colour image watermarking scheme resistant to affine geometric distortions. The technique involves embedding of the watermark in the host image in the Discrete Cosine Transform (DCT) domain in a spread-spectrum format. Harris corner detector based feature-points are then employed to get Delaunay triangulation which is used to reverse geometric attacks before attempting to extract the watermark. In some instances the geometric attack may lead to loss of feature-points resulting in the recovery of low quality watermark. In this paper we propose a scheme of estimating geometric attacks by taking the mean of selected triangles in the Delaunay tessellation. Computer simulation using MATLAB has been used to show the robustness of the proposed scheme.
  • Keywords
    discrete cosine transforms; feature extraction; image colour analysis; image watermarking; DCT domain; Delaunay tessellation; Delaunay triangulation; Harris corner detector based feature-points; Matlab; affine geometric attacks; affine geometric distortions; colour image watermarking technique scheme; discrete cosine transform; low quality watermark recovery; spread-spectrum format; watermark embedding; watermark extraction; Detectors; Discrete cosine transforms; Feature extraction; Image color analysis; Probes; Robustness; Watermarking; Delaunay triangulation; Discrete Cosine Transform; Image Watermarking and Vector Quantization;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    AFRICON, 2013
  • Conference_Location
    Pointe-Aux-Piments
  • ISSN
    2153-0025
  • Print_ISBN
    978-1-4673-5940-5
  • Type

    conf

  • DOI
    10.1109/AFRCON.2013.6757681
  • Filename
    6757681