• DocumentCode
    759962
  • Title

    Earmarked [biometrics]

  • Author

    Collins, Luke

  • Volume
    51
  • Issue
    11
  • fYear
    2005
  • Firstpage
    38
  • Lastpage
    40
  • Abstract
    Using ears as the basis for a biometric is a novel idea, but then ears have some novel features. They are normally found in the same place, on either side of the head, no matter who you are looking at. Many forms of biometrics are troubled by the ravages of time. The ear, however, has a rich and stable structure, although they may get bigger as the underlying cartilage grows, and is not affected by changes in facial expression. It is big, compared to a retina, iris or thumbprint, so it is easier to photograph. Its position on the side of the head makes the background predictable (hair or no hair), making distinguishing ´ear´ from ´not ear´ simpler. The fact that it is possible to create biometric data from a picture of an ear makes gathering that data more hygienic and less stressful than taking iris scans or fingerprints. These are some of the reasons why researchers at the University of Southampton have been investigating using ears to identify people uniquely over the long term. As with other forms of biometrics, simple images of an ear are not enough to identify one from another uniquely and reliably. Variations of the size and position of the ear within an image and of lighting conditions make direct comparisons unreliable. Instead, it is necessary to extract properties of the image that do not vary with these changes. The article examines the techniques used.
  • Keywords
    biometrics (access control); ear; feature extraction; image recognition; biometrics; cartilage; ears; facial expression; image processing; iris; person identification; property extraction; retina; thumbprint;
  • fLanguage
    English
  • Journal_Title
    IEE Review
  • Publisher
    iet
  • ISSN
    0953-5683
  • Type

    jour

  • DOI
    10.1049/ir:20051105
  • Filename
    1557558