95 articles 

inria-00104883, version 1

Modeling On-line Handwriting Using Pairwise Relational Features

Rudy Sicard () 1, Thierry Artières () 2, Eric Petit () 1

Tenth International Workshop on Frontiers in Handwriting Recognition (2006)

Abstract: We propose a new modeling strategy for on-line handwriting. It relies on the use of local and relational features. It allows implementing a variety of models, including traditional Markovian models. Introducing relational features allows building models that exhibit much robustness to noise, extra strokes, temporal ordering variations etc. It may be used for various tasks such as sequence recognition or partial matching of sequences.

  • 1:  France Télécom Recherche & Développement (FT R&D)
  • France Télécom
  • 2:  Laboratoire d'Informatique de Paris 6 (LIP6)
  • CNRS : UMR7606 – Université Pierre et Marie Curie (UPMC) - Paris VI
  • Domain : Computer Science/Document and Text Processing
    Computer Science/Computer Vision and Pattern Recognition
  • Keywords : Relational features – pairwise dependencies – robust segmentation – partial matching – diagnosis – recognition
  • Comment : http://www.suvisoft.com
 
  • inria-00104883, version 1
  • oai:hal.inria.fr:inria-00104883
  • From: 
  • Submitted on: Monday, 9 October 2006 16:02:01
  • Updated on: Monday, 9 October 2006 16:03:39