Incremental Construction of Neighborhood Graphs using the Ants Self-Assembly Behavior

Hanene Azzag (1) , Julien Lavergne (2) , Christiane Guinot (3) , Gilles Venturini (2)

1 LIPN - Laboratoire d'Informatique de Paris-Nord
2 LIFAT - Laboratoire d'Informatique Fondamentale et Appliquée de Tours
3 CERIES - CENTRE DE RECHERCHES ET D'INVESTIGATIONS EPIDERMIQUES ET SENSORIELLES

In this paper we present a new incremental algorithm for building neighborhood graphs between data. It is inspired from the self-assembly behavior observed in real ants where ants progressively become attached to an existing support and then successively to other attached ants. Each artifi- cial ant represents one data. The way ants move and build a graph depends on the similarity between the data. We have compared our results to those obtained by the relative neighborhood algorithm on several databases (either arti- ficial or real), and we show that our method is competitive especially with respect to execution times.

Domaines: Interface homme-machine [cs.HC]; Vision par ordinateur et reconnaissance de formes [cs.CV]

Hanene Azzag, Julien Lavergne, Christiane Guinot, Gilles Venturini. Incremental Construction of Neighborhood Graphs using the Ants Self-Assembly Behavior. 19th annual IEEE International Conference on Tools with Artificial Intelligence, 2007, Patrace, Greece. pp.399-406. ⟨hal-01024501⟩