Unsupervised Clustering on Multi-Components Datasets: Applications on Images and Astrophysics Data

Laurent Galluccio 1, * Olivier Michel 2 Pierre Comon 1
* Auteur correspondant
2 GIPSA-C2S - C2S
GIPSA-DIS - Département Images et Signal
Abstract : This paper proposes an original approach to cluster multi-component data sets with an estimation of the number of clusters. From the construction of a minimal spanning tree with Prim's algorithm and the assumption that the vertices are approximately distributed according to a Poisson distribution, the number of clusters is estimated by thresholding the Prim's trajectory. The corresponding cluster centroids are then computed in order to initialize the Generalized Lloyd's algorithm, also known as K-means, which allows to circumvent initialization problems. Metrics used for measuring similarity between multi-dimensional data points are based on symmetrical divergences. The use of these informational divergences together with the proposed method lead to better results than some other clustering methods in the framework of astrophysical data processing. An application of this method in the multi-spectral imagery domain with a satellite view of Paris is also presented.
Type de document :
Communication dans un congrès
16th European Signal Processing Conference (EUSIPCO-2008), Aug 2008, Lausanne, Switzerland. pp.P4-1, 2008
Liste complète des métadonnées

https://hal.archives-ouvertes.fr/hal-00339542
Contributeur : Laurent Galluccio <>
Soumis le : mardi 18 novembre 2008 - 11:01:15
Dernière modification le : mercredi 17 juin 2015 - 01:15:48
Document(s) archivé(s) le : lundi 7 juin 2010 - 20:42:53

Fichiers

GallMC08lausanne.pdf
Fichiers produits par l'(les) auteur(s)

Identifiants

  • HAL Id : hal-00339542, version 1

Collections

Citation

Laurent Galluccio, Olivier Michel, Pierre Comon. Unsupervised Clustering on Multi-Components Datasets: Applications on Images and Astrophysics Data. 16th European Signal Processing Conference (EUSIPCO-2008), Aug 2008, Lausanne, Switzerland. pp.P4-1, 2008. <hal-00339542>

Partager

Métriques

Consultations de
la notice

363

Téléchargements du document

104