• DocumentCode
    3278859
  • Title

    Caricaturation for human face pictures

  • Author

    Chang, I-Cheng ; Cheng, Ruei-min

  • Author_Institution
    Dept. of Comput. Sci. & Inf. Eng., Nat. Dong Hwa Univ., Hualien, Taiwan
  • Volume
    4
  • fYear
    2011
  • fDate
    10-13 July 2011
  • Firstpage
    1702
  • Lastpage
    1707
  • Abstract
    Being one of the attractive art forms, comic appears on many different media and is able to represent the subjects well with only a few strokes. The paper proposes a caricaturation system in which a face picture can be transformed into a caricatured output in accordance with the specific painter´s style. A facial feature rendering process and a hair rendering process are included in the system. We build a database in advance that stores the extracted information of the face pictures and caricatured face of a painting style. According to the input face picture, the proposed rendering processes compute the correspondence to the database so that it can select the appropriate caricatured facial features and hairs from the database, and adjust them to fit the input face. The proposed system can keep the characteristics of the input face as well as preserve the painter´s style. To increase the flexibility, an additional portrayal rendering process is used to add the shadows on the caricatured faces and feminize a male face.
  • Keywords
    face recognition; rendering (computer graphics); caricaturation; facial feature rendering process; hair rendering process; human face pictures; painting style; photorealistic image; portrayal rendering process; Databases; Face; Facial features; Hair; Mouth; Rendering (computer graphics); Shape; Nonphotorealistic rendering; example-based method; image caricaturation; image warping;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Machine Learning and Cybernetics (ICMLC), 2011 International Conference on
  • Conference_Location
    Guilin
  • ISSN
    2160-133X
  • Print_ISBN
    978-1-4577-0305-8
  • Type

    conf

  • DOI
    10.1109/ICMLC.2011.6017011
  • Filename
    6017011