• DocumentCode
    2075315
  • Title

    Efficient Generation of Shape-Based Reference Frames for the Corpus Callosum for DTI-based Connectivity Analysis

  • Author

    Sun, Hui ; Yushkevich, Paul A. ; Zhang, Hui ; Gee, James C. ; Simon, Tony J.

  • Author_Institution
    University of Pennsylvania, USA
  • fYear
    2006
  • fDate
    17-22 June 2006
  • Firstpage
    87
  • Lastpage
    87
  • Abstract
    Yushkevich et.al. [17, 18] established a PDE-based deformable modeling approach called continuous medial representation (cm-rep), in which the geometric relationship between the medial axis of a 3D object and its boundary is captured. Continuous medial description of an object not only provides useful shape features for object characterization and comparison; it also imposes a shape-based reference frame on the interior of that object. Such a reference frame provides a useful means of representing different instances of an anatomical structure using a common canonical parametrization domain. This paper presents an efficient method to construct continuous medial shape models for 2D objects. A closed form solution for the ordinary differential equation (ODE) is derived via Pythagorean hodograph (PH) curves. That closed form solution reduces the computation complexity from solving an ODE system to pure algebraic manipulation. Using this method, we generate shape-based reference frames, and demonstrate how they can be applied to the analysis of anatomical connectivity of corpora callosa, obtained by fiber tracking in diffusion tensor magnetic resonance imaging (DTI) in a chromosome 22q11.2 deletion syndrome study.
  • Keywords
    Anatomical structure; Biological cells; Closed-form solution; Deformable models; Differential equations; Diffusion tensor imaging; Image analysis; Magnetic analysis; Shape; Solid modeling;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Computer Vision and Pattern Recognition Workshop, 2006. CVPRW '06. Conference on
  • Print_ISBN
    0-7695-2646-2
  • Type

    conf

  • DOI
    10.1109/CVPRW.2006.68
  • Filename
    1640528