Network-based correlated correspondence for unsupervised domain adaptation of hyperspectral satellite images

Abstract : Adapting a model to changes in the data distribution is a relevant problem in machine learning and pattern recognition since such changes degrade the performances of classifiers trained on undistorted samples. This paper tackles the problem of domain adaptation in the context of hyperspectral satellite image analysis. We propose a new correlated correspondence algorithm based on network analysis. The algorithm finds a matching between two distributions, which preserves the geometrical and topological information of the corresponding graphs. We evaluate the performance of the algorithm on a shadow compensation problem in hyperspectral image analysis: the land use classification obtained with the compensated data is improved
Type de document :
Communication dans un congrès
International Conference on Pattern Recognition (ICPR 2014), Aug 2014, Stockholm, Sweden. pp.1--6, 2014
Liste complète des métadonnées

https://hal.archives-ouvertes.fr/hal-01018700
Contributeur : Nicolas Courty <>
Soumis le : vendredi 4 juillet 2014 - 16:50:25
Dernière modification le : jeudi 9 février 2017 - 16:04:30
Document(s) archivé(s) le : samedi 4 octobre 2014 - 13:00:47

Fichier

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

Identifiants

  • HAL Id : hal-01018700, version 1

Citation

Julien Rebetez, Devis Tuia, Nicolas Courty. Network-based correlated correspondence for unsupervised domain adaptation of hyperspectral satellite images. International Conference on Pattern Recognition (ICPR 2014), Aug 2014, Stockholm, Sweden. pp.1--6, 2014. <hal-01018700>

Partager

Métriques

Consultations de
la notice

203

Téléchargements du document

152