• DocumentCode
    3719674
  • Title

    A new definition of fuzzy contours in mammographic images

  • Author

    Marwa Hmida;Kamel Hamrouni;Basel Solaiman

  • Author_Institution
    Universite de Tunis El Manar, Ecole Nationale d´Ingenieurs de Tunis, LR-SITI Signal Image et Technologie de reformation, Tunis, Tunisie
  • fYear
    2015
  • Firstpage
    169
  • Lastpage
    173
  • Abstract
    In this paper, we give a new definition of fuzzy contours in mammographic images and we propose an approach to get them. This approach starts from region of interest images with initial contours delimiting masses provided by an expert radiologist, and then it involves three major stages: mass segmentation using geometric deformable models, contour representation and finally definition of fuzzy contours. For the first phase, the initial zero level set is performed by using the initial contour guided by the radiologist. After that, we propose a new representation of contours. Since the segmentation of masses in mammograms is a difficult problem due to the characteristics of mammographic image, and even mass edges can be identified differently by different radiologists and also by the same radiologist at a different time, we define an entire region that may contain the accurate edges, in which we assign to each pixel a membership value to the class "Contour". We call that fuzzy contours.
  • Keywords
    "Decision support systems","Image segmentation","Image edge detection","Yttrium","Deformable models","Level set"
  • Publisher
    ieee
  • Conference_Titel
    Image Processing Theory, Tools and Applications (IPTA), 2015 International Conference on
  • Print_ISBN
    978-1-4799-8636-1
  • Electronic_ISBN
    2154-512X
  • Type

    conf

  • DOI
    10.1109/IPTA.2015.7367120
  • Filename
    7367120