Modeling Relative Positioning of Handwritten Patterns
Résumé
In this paper we present a new generic method for modeling and evaluating the relative positioning of handwritten patterns. Thanks to a morphological approach it handles the variability of handwritten shapes and providing a flexible fuzzy description of positioning relations. We propose a way to automatically learn such spatial relation models from samples, and prove that resulting models can be fruitfully used for pattern recognition purpose.
| Origine | Fichiers produits par l'(les) auteur(s) |
|---|
Loading...