|
50: Digital Geometry Processing
Tuesday, Full Day, 8:30 am - 5 pm
Room 502A
Traditionally,
fine-detail geometry is represented through unstructured
polygonal meshes, which are awkward for editing, filtering,
and compression applications. This course proposes a new
paradigm based on semi-regular meshes constructed through
a process of recursive quadrisection, which, according
to recent research, has many advantages. Presenters show
how to build semi-regular meshes from unstructured polygonal
meshes and raw range data, and how to build applications
such as filtering, editing, simulation, and compression
using semi-regular meshes.
Prerequisites
Familiarity with basic graphics algorithms. Some prior
exposure to polygonal meshes.
Topics
3D acquisition: geometry reconstruction, denoising, appearance
acquisition, multi-view integration. Semi-regular meshes:
definition, properties, inherent advantages, construction,
normal meshes, conversion from other formats. Subdivision
and details as a natural replacement for classical Fourier
analysis to arbitrary topology surfaces. Hierarchical
editing: applications of digital geometry, processing
algorithms in interactive modeling and physical simulation.
Simulation: Use of multi-resolution geometry in physical
modeling for animation, engineering design, and scientific
computing. Compression: wavelets on semi-regular meshes,
lifting, progressivity, zero-tree coding, quantization,
rate-distortion trade-offs.
Organizers
Peter Schröder
California Institute of Technology
Wim Sweldens
Lucent Technologies, Bell Laboratories
Lecturers
Brian Curless
University of Washington
Igor Guskov
California Institute of Technology
Peter Schröder
California Institute of Technology
Wim Sweldens
Lucent Technologies, Bell Laboratories
Denis Zorin
New York University |
|
