Share Email Print
cover

Proceedings Paper

Combined SPHARM-PDM and entropy-based particle systems shape analysis framework
Author(s): Beatriz Paniagua; Lucile Bompard; Josh Cates; Ross Whitaker; Manasi Datar; Clement Vachet; Martin Styner
Format Member Price Non-Member Price
PDF $14.40 $18.00

Paper Abstract

Description of purpose: The NA-MIC SPHARM-PDM Toolbox represents an automated set of tools for the computation of 3D structural statistical shape analysis. SPHARM-PDM solves the correspondence problem by defining a first order ellipsoid aligned, uniform spherical parameterization for each object with correspondence established at equivalently parameterized points. However, SPHARM correspondence has shown to be inadequate for some biological shapes that are not well described by a uniform spherical parameterization. Entropy-based particle systems compute correspondence by representing surfaces as discrete point sets that does not rely on any inherent parameterization. However, they are sensitive to initialization and have little ability to recover from initial errors. By combining both methodologies we compute reliable correspondences in topologically challenging biological shapes. Data: Diverse subcortical structures cohorts were used, obtained from MR brain images. Method(s): The SPHARM-PDM shape analysis toolbox was used to compute point based correspondent models that were then used as initializing particles for the entropy-based particle systems. The combined framework was implemented as a stand-alone Slicer3 module, which works as an end-to-end shape analysis module. Results: The combined SPHARM-PDM-Particle framework has demonstrated to improve correspondence in the example dataset over the conventional SPHARM-PDM toolbox. Conclusions: The work presented in this paper demonstrates a two-sided improvement for the scientific community, being able to 1) find good correspondences among spherically topological shapes, that can be used in many morphometry studies 2) offer an end-to-end solution that will facilitate the access to shape analysis framework to users without computer expertise.

Paper Details

Date Published: 13 April 2012
PDF: 7 pages
Proc. SPIE 8317, Medical Imaging 2012: Biomedical Applications in Molecular, Structural, and Functional Imaging, 83170L (13 April 2012); doi: 10.1117/12.911228
Show Author Affiliations
Beatriz Paniagua, The Univ. of North Carolina at Chapel Hill (United States)
Lucile Bompard, The Univ. of North Carolina at Chapel Hill (United States)
Josh Cates, Univ. of Utah (United States)
Ross Whitaker, Univ. of Utah (United States)
Manasi Datar, Univ. of Utah (United States)
Clement Vachet, The Univ. of North Carolina at Chapel Hill (United States)
Martin Styner, The Univ. of North Carolina at Chapel Hill (United States)


Published in SPIE Proceedings Vol. 8317:
Medical Imaging 2012: Biomedical Applications in Molecular, Structural, and Functional Imaging
Robert C. Molthen; John B. Weaver, Editor(s)

© SPIE. Terms of Use
Back to Top