Mémoire visuelle pour l'analyse de documents structurés

Joseph Chazalon 1 Bertrand Coüasnon 1 Aurélie Lemaitre 1
1 IMADOC - Interprétation et Reconnaissance d’Images et de Documents
UR1 - Université de Rennes 1, INSA Rennes - Institut National des Sciences Appliquées - Rennes, CNRS - Centre National de la Recherche Scientifique : UMR6074
Abstract : Current analysis methods propose systems which are adapted to recognize sets of documents of a same kind (statistical methods, grammatical analysis), but each page is then processed in isolation. However, in a collection context, it is important to make the most of the result of the processing of one or several pages in order to improve the later processing of other pages. Thus, we propose the concept of visual memory so as to enable existing recognition systems to use, locally and at run time, contextual information for each page. We detail its specifications which allow an implementation in any recognition system. Finally, we present our implementation using a grammatical method, and its application for several types of information flow.
Document type :
Conference papers
Complete list of metadatas

https://hal.inria.fr/inria-00545886
Contributor : Joseph Chazalon <>
Submitted on : Monday, December 13, 2010 - 11:04:18 AM
Last modification on : Friday, November 16, 2018 - 1:28:14 AM
Long-term archiving on : Monday, November 5, 2012 - 1:21:09 PM

Identifiers

  • HAL Id : inria-00545886, version 1

Citation

Joseph Chazalon, Bertrand Coüasnon, Aurélie Lemaitre. Mémoire visuelle pour l'analyse de documents structurés. Colloque International Francophone sur l'Ecrit et le Document (CIFED2010), Mar 2010, Sousse, Tunisie. ⟨inria-00545886⟩

Share

Metrics

Record views

205

Files downloads

240