6.838/4.214:
Interactive Geometric Data Structures and Computation
Meeting 26: Course Summary
Last
Meeting
Talk tomorrow:
Pat
Hanrahan, 330pm in 34-101 (refreshments at 315)
Digital
Lights, Cameras, and Materials
Final project presentations:
40 minute
slots, arrange via e-mail to bryt@graphics
Some big questions in geometric
computing:
-
Capture
-
Representation
-
Inspection
-
Responsiveness
-
Simulation
-
Design
-
Visual inspection
What we covered:
-
Representation I
-
Geometric primitives
-
Operators (sidedness, etc.)
-
Constructions vs. Closed-form
expressions
-
Duality!
-
... of geometric primitives
-
... of topological structure
-
Adjacency data structures
-
Winged-edge, quad-edge, etc.
-
Workhorse algorithms
-
Convex hull (2D, 3D, general
d)
-
Delaunay Triangulation/Voronoi
Diagram
-
Quadtree/Octree/k-D tree
-
Linear programming
-
Arrangements
-
Representation II
-
3D surfaces/solids
-
Implicit, boundary, discretized
-
Parametric curves and surfaces
-
Lifting transforms, geometric
reductions
-
Voronoi diagram (d-dim) to
convex hull (d+1 - dim
-
Using existing algorithm to
solve a new problem
-
Degeneracy !
-
Notion of general position
(generic input)
-
Robust predicates: more
precision, SoS, ...
-
Hybrid data structures
-
Arrangements of parametric
curves/surfaces
-
Algorithmic paradigms
-
Brute-force, divide & conquer,
incremental, sweep-line/plane
-
Kinetic framework
-
Iterative algorithms
-
Estimating camera positions,
features
-
Estimating point/line nearest
to k 3D lines
-
Applications
-
Surface reconstruction from
volume data (marching cubes
-
Visibility (aspect graph, 3D
linespace)
-
Radiance field capture (3D linespace)
-
Semi-automatic IK skeletons
(3D voronoi, medial axis)
-
Architectural visibility (BSP
trees, constrained search)
-
3D reconstruction from imagery
(robust stats, sweep algs)
Some things we didn't cover
(Sources Eppstein, Erickson, Amenta)
-
Interpolation by curves, surfaces
-
... given some number of samples
of a function,
estimate the function at some other points
-
Ideas ?
-
Subdivision surfaces
-
Curved surface generation through
subdivision & averaging
-
Terrain and GIS (geographic
information systems)
-
Data with attributes (income,
race, children...)
-
External memory algorithms/Adaptivity/Responsiveness
-
Mesh generation
-
CFD (air flow, etc.)
-
Constrained tetrahedralization/hexahedralization
-
Even harder: do so with bounded
aspect ratio!
-
Parallelization
-
Minimum spanning trees
-
Cosmology, organic systems,
network design...
-
Assembly planning
-
Motion planning, grasping,
assembly sequences
-
CAD
-
Interference detection, LOD,
interaction & editing
-
Intuitive 3D modeling
operations
-
CAM
-
Tool sequencing, material use
optimization
-
Protein folding
-
What shape does a given base-pair
sequence assume?
-
Computational drug design
-
design molecule to bind to
a given receptor
-
VLSI design
-
Routing; dissipation; 3D structures
-
Graph drawing
-
Genealogy, power networks,
dynamic systems
-
Alignment
-
... of DNA sequences
-
... of multiple 3D observations
(MRI, range data, etc.)
-
Medical Imaging
-
3D reconstruction from voxels,
slices, etc.
Directions in geometric computing
-
Capture via sensors
-
Cameras, range-sensors, MRI
...
-
Information hiding
-
Robustness
-
Self-checking codes; abstraction
-
Responsiveness in the face
of complexity
-
Efficient data structures;
external memory
-
Responsiveness in the face
of change
-
Continuous / discrete hybrids
-
Physically-based simulation
-
Pedagogy!
-
Visual inspection, exposition
Summary
-
Course introduced a collection
of tools & techniques for
-
Representation (analytic/approximate)
-
Construction & Maintenance
-
Inspection & Simulation
-
Maintaining responsiveness
-
Application building
-
Visual, analytic inspection
Last Meeting .... Next
Meeting .... Course
Page .... Meeting Schedule
Created: Feb 1998
Prof.
Seth Teller, MIT Computer Graphics Group, teller@graphics.lcs.mit.edu