Qualitative Geometry for Shape Recognition

  • Authors:
  • Vincent Dugat;Pierre Gambarotto;Yannick Larvor

  • Affiliations:
  • Institut de Recherche en Informatique de Toulouse (IRIT), 118 Route de Nárbonne, Toulouse Cedexy, France F-31062. dugat@cict.fr;Institut de Recherche en Informatique de Toulouse (IRIT), 118 Route de Nárbonne, Toulouse Cedexy, France F-31062. gamba@irit.fr;Institut de Recherche en Informatique de Toulouse (IRIT), 118 Route de Nárbonne, Toulouse Cedexy, France F-31062. larvor@irit.fr

  • Venue:
  • Applied Intelligence
  • Year:
  • 2002

Quantified Score

Hi-index 0.01

Visualization

Abstract

We propose an algorithm providing an abstract representation of any polygonal object iO in terms of spheres. The result is a graph-based skeleton capturing the general shape of iO and its inner structure (respective positions of convex parts and their thickness). We define a first-order logic language expressing in a qualitative way the needed notions (distance, size and angle). Last, we propose methods to compare shapes using this graph-based skeleton of objects.