• DocumentCode
    525632
  • Title

    Adaptive tables based fast B-spline transformation for facial images deformation

  • Author

    Ohmura, Mizuho ; Seo, Masataka ; Aoki, Hiromatsu ; Chen, Yen-wei

  • Author_Institution
    Grad. Sch. of Sci. & Eng., Ritsumeikan Univ., Kusatsu, Japan
  • fYear
    2010
  • fDate
    23-25 June 2010
  • Firstpage
    640
  • Lastpage
    643
  • Abstract
    Photo editing techniques are getting more and more attention, especially the function of face deformation for beauty, with the popularization of mobile phone cameras. This paper focuses on the development of face deformation algorithm to make the face in the photo look more attractive, which can be used in the photo editor of mobile phone. A fast B-spline transformation method is proposed for face deformation or warping. The proposed method is based on adaptive tables and can achieve the same accuracy with the conventional method, but can largely lower the computational cost. In the conventional image deformation methods, large numbers of face feature points are needed for the transform parameter calculation, so they are computationally expensive. However, in the proposed method, only two points are needed. Furthermore, an adaptive table holding the standard facial quantities is employed to reduce the computing time. The experiments showed an encourage result, which indicate that it is an practical technique for face beauty in mobile phone.
  • Keywords
    cameras; deformation; face recognition; mobile handsets; splines (mathematics); computational cost; face deformation algorithm; facial images deformation; fast B-spline transformation adaptive tables; mobile phone cameras; photo editing techniques; transform parameter calculation; warping; Decision support systems; Spline; Virtual reality; Adaptive Tables; B-spline transformation; face beauty; image deformation;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Software Engineering and Data Mining (SEDM), 2010 2nd International Conference on
  • Conference_Location
    Chengdu
  • Print_ISBN
    978-1-4244-7324-3
  • Electronic_ISBN
    978-89-88678-22-0
  • Type

    conf

  • Filename
    5542846