• DocumentCode
    1860290
  • Title

    Image harmony for consumer images

  • Author

    Fedorovskaya, Elena ; Neustaedter, Carman ; Hao, Wei

  • Author_Institution
    Res. Labs., Eastman Kodak Co., Rochester, NY
  • fYear
    2008
  • fDate
    12-15 Oct. 2008
  • Firstpage
    121
  • Lastpage
    124
  • Abstract
    Images elicit a variety of emotional responses related to image content, overall aesthetic appeal, or a combination of both. One aspect of aesthetic appeal is harmony: the pleasing or congruent arrangement of parts producing internal calm or tranquility. We conducted a series of experiments to identify what low level features could predict harmony in an image. Subjective judgments of image harmony were collected for images representative of typical consumer photography. Our initial results show that for simplified images (pixelated to control for emotional responses to scenes and objects) impressions of image harmony depend on statistical properties of low level local features. The feature combinations vary for different individuals but typically involve edge contrast, average lightness and range of lightness. At the same time inclusion of Gestalt principles is needed to account for the subjects´ data. Additionally, global low level features related to spatial image structure may help to explain results with black and white and color images.
  • Keywords
    image colour analysis; image representation; Gestalt principles; aesthetic appeal; black and white images; color images; consumer images; consumer photography; image content; image harmony; low level features; statistical properties; Animals; Art; Color; Dictionaries; Digital images; Laboratories; Layout; Photography; Pixel; Psychology; computational assessment; consumer photography; image harmony; image quality;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Image Processing, 2008. ICIP 2008. 15th IEEE International Conference on
  • Conference_Location
    San Diego, CA
  • ISSN
    1522-4880
  • Print_ISBN
    978-1-4244-1765-0
  • Electronic_ISBN
    1522-4880
  • Type

    conf

  • DOI
    10.1109/ICIP.2008.4711706
  • Filename
    4711706