• DocumentCode
    3484203
  • Title

    Ultrasound despeckling for active contour segmentation

  • Author

    Ma, Yiming ; Tay, Peter C.

  • Author_Institution
    Dept. of Electr. & Comput. Eng. Technol., Western Carolina Univ., Cullowhee, NC, USA
  • fYear
    2009
  • fDate
    7-10 Nov. 2009
  • Firstpage
    3357
  • Lastpage
    3360
  • Abstract
    Automatic or semi-automatic gradient based segmentation of ultrasound images is problematic because the inherent nature of ultrasound imaging produces speckle. The main concern of this paper is to remove speckle cause gradients that cannot be attributed to anatomic features. Application of despeckling algorithms attempt to remove or reduce the speckle noise and thus the pixelwise gradients caused by speckle. Yet, the despeckling algorithm should preserve edges and gradients that are due to anatomic features such as the boundaries of organs, differing tissues, etc. This paper proposes the use of the noise enhanced squeeze box filter to despeckle cardiac ultrasound images so that an application of well-know gradient based active contour methods can be used for segmentation.
  • Keywords
    filtering theory; gradient methods; image denoising; image segmentation; medical image processing; active contour segmentation; cardiac ultrasound image despeckling; gradient based active contour method; noise enhanced squeeze box filter; semiautomatic gradient based segmentation; ultrasound despeckling algorithm; ultrasound imaging; Active contours; Active noise reduction; Filtering; Filters; Image edge detection; Image segmentation; Iterative algorithms; Noise robustness; Speckle; Ultrasonic imaging; Ultrasound imaging; cardiac segmentation; despeckling;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Image Processing (ICIP), 2009 16th IEEE International Conference on
  • Conference_Location
    Cairo
  • ISSN
    1522-4880
  • Print_ISBN
    978-1-4244-5653-6
  • Electronic_ISBN
    1522-4880
  • Type

    conf

  • DOI
    10.1109/ICIP.2009.5413890
  • Filename
    5413890