Interpolation and Intersection Algorithms and GPU
Date issued
2012
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
IARIA
Abstract
Interpolation and intersection methods are closely
related and used in computer graphics, visualization, computer
vision etc. The Euclidean representation is used nearly exclusively
not only in computational methods, but also in education despite it
might lead to instability in computation in many cases. The
projective geometry, resp. projective extension of the Euclidean
space, offers many positive features from the computational and
educational points of view with higher robustness and stability of
computation. This paper presents simple examples of projective
representation advantages, especially from the educational point of
view. In particular, how interpolation and intersection can be
applied to fundamental algorithms, which are becoming more
robust, stable and faster due to compact formulation. Another
advantage of the proposed approach is a simple implementation on
vector-vector architectures, e.g. GPU, as it is based on matrixvector
operations.
Description
Subject(s)
interpolační algoritmy, algoritmy průsečíků, princip duality, barycentrické souřadnice, lineární systém rovnic
Citation
ICONS 2012, Saint Gilles, Reunion Island, p. 193-198.