Skip to Main content Skip to Navigation
Theses

Approche markovienne bidimensionnelle d'analyse et de reconnaissance de documents manuscrits

Abstract : This thesis presents a 2D generic stochastic approach for the analysis and recognition of handwritten documents, called AMBRES (Approche Markovienne Bidimensionnelle pour la Reconnaissance et la Segmentation d'images), which relies on Markov random fields, 2D dynamic programming and 2D image analysis.

AMBRES has been successfully applied to such various tasks as handwritten character / word recognition, handwritten document structuring and logo recognition, and could be extended to other computer vision issues.

Rigorous protocols have been used for characterizing the system, studying its parameters and assessing its performances. In particular, AMBRES has been assessed within the framework of the French RIMES (Reconnaissance et Indexation de données Manuscrites et de fac similES) evaluation campaign.
Document type :
Theses
Complete list of metadatas

Cited literature [97 references]  Display  Hide  Download

https://tel.archives-ouvertes.fr/tel-00273255
Contributor : Francoise Preteux <>
Submitted on : Monday, April 14, 2008 - 5:48:13 PM
Last modification on : Friday, October 23, 2020 - 4:39:38 PM
Long-term archiving on: : Friday, September 28, 2012 - 12:35:48 PM

Identifiers

  • HAL Id : tel-00273255, version 1

Citation

Melanie Lemaitre. Approche markovienne bidimensionnelle d'analyse et de reconnaissance de documents manuscrits. Informatique [cs]. Université René Descartes - Paris V, 2007. Français. ⟨tel-00273255⟩

Share

Metrics

Record views

344

Files downloads

1051