Attribute Profiles from Partitioning Trees

Petra Bosilj (1, 2) , Bharath Bhushan Damodaran (1) , Erchan Aptoula (3) , Mauro Dalla Mura (4) , Sébastien Lefèvre (1)

1 OBELIX - Environment observation with complex imagery
2 University of Lincoln
3 Gebze Teknik Üniversitesi [Gebze]
4 GIPSA-SIGMAPHY - GIPSA - Signal Images Physique

Morphological attribute profiles are among the most prominent spatial-spectral pixel description tools. They can be calculated efficiently from tree based representations of an image. Although widely and successfully used with various inclusion trees (i.e., component trees and tree of shape), in this paper, we investigate their implementation through partitioning trees, and specifically α-and (ω)-trees. Our preliminary findings show that they are capable of comparable results to the state-of-the-art, while possessing additional properties rendering them suitable for the analysis of multivariate images.

Mots clés: attribute profiles; partitioning trees; α-tree; (ω)-tree; hy- perspectral images

Domaines: Traitement des images [eess.IV]

Petra Bosilj, Bharath Bhushan Damodaran, Erchan Aptoula, Mauro Dalla Mura, Sébastien Lefèvre. Attribute Profiles from Partitioning Trees. ISMM 2017 - 13th International Symposium on Mathematical Morphology, May 2017, Fontainebleau, France. ⟨10.1007/978-3-319-57240-6_31⟩. ⟨hal-01672856⟩