• DocumentCode
    322889
  • Title

    Lip motion capture and its application to 3-D molding

  • Author

    Okubo, Masashi ; Watanabe, Tomio

  • Author_Institution
    Okayama Prefectural Univ., Japan
  • fYear
    1998
  • fDate
    14-16 Apr 1998
  • Firstpage
    187
  • Lastpage
    192
  • Abstract
    It is essential for machine lip reading to process not only static images but also moving images. This paper proposes a method, which is named Optical Snakes, for lip motion extraction from moving images using the cooperation between the active contour models called Snakes and the optical flow. The effectiveness of the method is demonstrated by the extraction of a series of lip contours from time-varying images without any markers, and by the analysis of lip contours for machine lip reading. Some real products which characterize a speaker´s pronunciation are made from the lip motion data by a photoforming system. It is demonstrated that the real products are better than the virtual products on a computer display for the evaluation of shapes. The application of lip motion to molding would lead to a new method for the analysis of lip motion and the training of pronunciation
  • Keywords
    CAD; face recognition; feature extraction; image sequences; motion estimation; user interfaces; 3D molding; Optical Snakes; active contour models; computer display; lip contours; lip motion capture; lip motion extraction; machine lip reading; moving images; optical flow; photoforming system; pronunciation; static images; three dimensional molding; time-varying images; training; Active contours; Active noise reduction; Humans; Image analysis; Image motion analysis; Motion analysis; Noise measurement; Petroleum; Shape; Spline;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Automatic Face and Gesture Recognition, 1998. Proceedings. Third IEEE International Conference on
  • Conference_Location
    Nara
  • Print_ISBN
    0-8186-8344-9
  • Type

    conf

  • DOI
    10.1109/AFGR.1998.670947
  • Filename
    670947