On morphological color texture characterization
Résumé
We investigate the combined use of multiple structuring elements with the standard morphological texture characterization tools, namely morphological covariance and granulometry. The resulting operator is applied to both grayscale and color images in the context of texture classication. As to its extension to color texture data, it is realized by means of a weighting based reduced vector ordering in the IHLS color space, equipped with genetically optimized arguments. The classication experiments based on this framework are carried out with the publically available Outex13 texture database, where the proposed feature extraction scheme outperforms the uni-variable versions of the operators under consideration.
Domaines
Traitement des images [eess.IV]Origine | Fichiers produits par l'(les) auteur(s) |
---|
Loading...