Department of Mathematics
 Search | Help | Login | pdf version | printable version

Math @ Duke



Publications [#265044] of Guillermo Sapiro

Papers Published

  1. Wirth, B; Bar, L; Rumpf, M; Sapiro, G, Geodesics in shape space via variational time discretization, Lecture Notes in Computer Science (Including Subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 5681 LNCS (November, 2009), pp. 288-302, Springer Berlin Heidelberg, ISSN 0302-9743 [doi]
    (last updated on 2019/06/26)

    A variational approach to defining geodesics in the space of implicitly described shapes is introduced in this paper. The proposed framework is based on the time discretization of a geodesic path as a sequence of pairwise matching problems, which is strictly invariant with respect to rigid body motions and ensures a 1-1 property of the induced flow in shape space. For decreasing time step size, the proposed model leads to the minimization of the actual geodesic length, where the Hessian of the pairwise matching energy reflects the chosen Riemannian metric on the shape space. Considering shapes as boundary contours, the proposed shape metric is identical to a physical dissipation in a viscous fluid model of optimal transportation. If the pairwise shape correspondence is replaced by the volume of the shape mismatch as a penalty functional, for decreasing time step size one obtains an additional optical flow term controlling the transport of the shape by the underlying motion field. The implementation of the proposed approach is based on a level set representation of shapes, which allows topological transitions along the geodesic path. For the spatial discretization a finite element approximation is employed both for the pairwise deformations and for the level set representation. The numerical relaxation of the energy is performed via an efficient multi-scale procedure in space and time. Examples for 2D and 3D shapes underline the effectiveness and robustness of the proposed approach. © 2009 Springer.
ph: 919.660.2800
fax: 919.660.2821

Mathematics Department
Duke University, Box 90320
Durham, NC 27708-0320