Zoning and Metaclasses Improving the Character Recognition

Abstract : The contribution of this paper is twofold. First we investigate the use of the confusion matrices in order to get some insight to better define perceptual zoning for character recognition. We will see that this idea can be used as a tool to support the design of zoning. The features considered in this work are based on concavities/convexities deficiencies, which are obtained by labeling the background pixels of the input image. Four different perceptual zoning (symmetrical and nonsymmetrical) are discussed. Experiments show that this mechanism of zoning could be considered as a reasonable alternative to exhaustive search algorithms. The second contribution is a methodology to define metaclasses for the problem of handwritten character recognition. The proposed approach is based on the disagreement among the characters and it uses Euclidean distance computed between the confusion matrices. Through comprehensive experiments we demonstrate that the use of meta-classes can improve the performance of the character recognition system.
Type de document :
Communication dans un congrès
Guy Lorette. Tenth International Workshop on Frontiers in Handwriting Recognition, Oct 2006, La Baule (France), Suvisoft, 2006
Liste complète des métadonnées

Littérature citée [15 références]  Voir  Masquer  Télécharger

https://hal.inria.fr/inria-00108305
Contributeur : Ist Rennes <>
Soumis le : vendredi 20 octobre 2006 - 13:12:00
Dernière modification le : lundi 20 juin 2016 - 14:10:32
Document(s) archivé(s) le : mardi 6 avril 2010 - 20:16:49

Identifiants

  • HAL Id : inria-00108305, version 1

Collections

Citation

Cinthia O. de A. Freitas, Luiz S. Oliveira, Edson Justino, Simone B. K. Aires. Zoning and Metaclasses Improving the Character Recognition. Guy Lorette. Tenth International Workshop on Frontiers in Handwriting Recognition, Oct 2006, La Baule (France), Suvisoft, 2006. 〈inria-00108305〉

Partager

Métriques

Consultations de la notice

92

Téléchargements de fichiers

113