Multi-scale Analysis of Uncertain Data: the λ-complex Filtration

Tom Dreyfus (Équipe ABS, Inria, Sophia-Antipolis)

The (weighted) α-complex allows one to perform a multi-scale analysis of a collection of points (balls), in conjunction with a growth model encoded in Voronoi (power) diagrams. α-complexes are thus associated to affine Voronoi diagrams, and have been instrumental in surface reconstruction, geometric inference, as well as molecular modeling.

This work is concerned with the 3D λ-complex, namely the equivalent of the α-complex for an additively-multiplicatively weighted distance. Such a distance is associated with so-called compoundly weighted Voronoi diagrams, whose bissectors are degree four algebraic surfaces. We present the filtration associated with the λ-complex, together with the predicates required to compute it. Finally, we discuss an application to represent toleranced i.e. uncertain models in molecular modeling.


Joint work with Frédéric Cazals.


This document was translated from LATEX by HEVEA.