• DocumentCode
    889611
  • Title

    Digital analysis of rotated images

  • Author

    Alliney, Stefano

  • Author_Institution
    Istituto di Matematica Gen. e Finanziaria, Bologna Univ., Italy
  • Volume
    15
  • Issue
    5
  • fYear
    1993
  • fDate
    5/1/1993 12:00:00 AM
  • Firstpage
    499
  • Lastpage
    504
  • Abstract
    The problem of estimating the relative rotations of two images taken from the same scene is considered, and a numerical method for computing such rotations is presented. The main features of the proposed method consist of the possibility of fast in-place computations with no preliminary interpolation of the data and a considerable precision for moderate rotation angles. The final estimate is performed with a least square criterion, but in the present case, even such computations can be substantially simplified. A detailed analysis shows that for practical applications, the influence of discretization errors does not affect, in a relevant way, the precision of the results; that is also confirmed by numerical experiments
  • Keywords
    image processing; least squares approximations; motion estimation; image processing; least square criterion; motion estimation; numerical method; relative rotation estimation; rotated images; Biomedical computing; Correlation; Density functional theory; Image analysis; Image motion analysis; Image registration; Interpolation; Layout; Least squares approximation; Motion estimation;
  • fLanguage
    English
  • Journal_Title
    Pattern Analysis and Machine Intelligence, IEEE Transactions on
  • Publisher
    ieee
  • ISSN
    0162-8828
  • Type

    jour

  • DOI
    10.1109/34.211470
  • Filename
    211470