• DocumentCode
    3579773
  • Title

    Moment Invariants Based on SUSAN Initial Edge Response

  • Author

    Guo Ping Xiong ; Jian Gang Tang

  • Author_Institution
    Shanghai Med. Instrum. Coll., Shanghai, China
  • Volume
    1
  • fYear
    2014
  • Firstpage
    22
  • Lastpage
    25
  • Abstract
    In view of problem that it usually lacks of fusion depth to combine moment invariants with other extraction technology on local invariant features, this article proposes the moment invariants based on SUSAN initial edge response, which provides a sort of organic integration for moment invariants and SUSAN edge detection algorithm. The combination point is to treat SUSAN initial edge response matrix as gray scale matrix, then to calculate its corresponding moment invariants. Experiments show that the recall and precision of retrieval algorithm based on the above moment invariants will become satisfied as soon as the quality of retrieval results reaches to a critical value, and the moment invariants reflect certain invariance about affine transformation and illumination change.
  • Keywords
    edge detection; feature extraction; image fusion; image retrieval; matrix algebra; SUSAN edge detection algorithm; SUSAN initial edge response matrix; affine transformation; fusion depth; gray scale matrix; illumination change; local invariant feature extraction technology; moment invariants; retrieval algorithm; smallest univalue segment assimilating nucleus; Algorithm design and analysis; Brightness; Feature extraction; Image edge detection; Image retrieval; Noise; (SUSAN) smallest univalue segment assimilating nucleus; (USAN) univalue segment assimilating nucleus; global invariant features; image retrieval; initial edge response; local invariant features; moment invariants;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Computational Intelligence and Design (ISCID), 2014 Seventh International Symposium on
  • Print_ISBN
    978-1-4799-7004-9
  • Type

    conf

  • DOI
    10.1109/ISCID.2014.20
  • Filename
    7064070