6: Object Space Visibility
Sunday, Tutorial, 10 am  noon
Room 406
A
survey of object space since the first algorithm was introduced
in 1963. In recent years, objectspace computation has
been considered an esoteric topic of only theoretical
interest. But objectspace algorithms solve problems,
such as the ability to arbitrarily magnify output images,
that otherwise require esoteric knowledge using rasterbased
techniques. By showing a number of problems and their
relevance to mainstream and fundamental problems in computer
graphics this tutorial stimulates revived interest in
objectspace computation.
Prerequisites
Familiarity with standard visibility techniques in computer
graphics such as scanline rendering and binary space partitioning.
Completion of the equivalent of a thirdyear course on
analysis of algorithms in a standard computer science
curriculum. Also helpful: a basic foundation in objectoriented
terminology and concepts.
Topics
Definition of and the need of objectspace computation.
Visibility in flatland (in 2D). Visibility in space compared
with visibility in the plane. Lines, conics, and quadrics
in Euclidean space. Survey of algorithms for visible line
determination, visible surface determination, and computation
of the visibility map. Performance gap between practical
and theoretical visibility algorithms. Research relevance:
meshing for radiosity.
Organizer/Lecturer
Sherif Ghali
MaxPlanckInstitut für Informatik 
