Dynamic Particles Systems for Visualization, Meshing, and Correspondence
Speaker: Ross T Whitaker , University of Utah/SCI Institute
Date: October 5 2007
Time: 2:30PM to 3:30PM
Location: 32-D507
Host: Polina Golland, CSAIL
Contact: Polina Golland, x38005, polina@csail.mit.edu
Relevant URL: The problem of representing continuous surfaces with discrete
geometric primitives is ubiquitous in a variety of fields including
engineering design, graphics, visualization, and computer vision. The
placement of points (the sampling problem) and their topological
relationships (the meshing problem) are intertwined. Yet point
placement and connectivity must satisfy a variety of interrelated and
often competing criteria, such as geometric and topological
regularity, accuracy, complexity, and adaptivity. This talk presents a
methodology for distributing point samples, or particles, on surfaces
by minimizing potential functions, with elliptic behaviors, in order
to achieve point distributions that satisfy a range of important
criteria. We describe an application of this technology to
visualizing isosurfaces of functions defined on curvilinear spaces,
which occur for instance, in simulations that result from high-order
finite-element schemes. We then present a pipeline for surface
meshing, which controls particle density using a smooth sizing field
in order to produce high-quality, topologically correct, triangle
meshes. This method generalizes to multiple materials (or labels) to
produce nonmanifold meshes that explicitly capture discontinuities at
the junctions were multiple materials meet. Finally we describe a
framework in which point samples are positioned across an ensemble of
shapes in order to minimize the information content of the resulting
correspondences. We show applications of these correspondences to
problems in the analysis of biological shapes.
See other events that are part of Biomedical Imaging and Analysis 2007/2008
See other events happening in October 2007
