A Priori and A Posteriori Integration and Combination of Language Models in an On-line Handwritten Sentence Recognition System

Solen Quiniou 1 Eric Anquetil 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 : This paper investigates the integration of different language models into an on-line sentence recognition system. The impact of n-gram and n-class (based on statistically and on morpho-syntactically classes) models, built on the Brown corpus, is compared in terms of word recognition rate. Furthermore, their integration in different steps of the recognition process (during it or to rescore the Nbest list of proposed sentences) is considered, thus showing better performances when used the sooner. Combinations of these models are also studied, in addition to the integration in the aforementioned recognition steps. All experiments are carried out on sentences from the Brown corpus which were written by several writers.
Keywords : N-best list rescoring. statistical language models On-line sentence recognition
Document type :
Conference papers
Complete list of metadatas

Cited literature [18 references]  Display  Hide  Download
https://hal.inria.fr/inria-00105112
Contributor : Anne Jaigu <>
Submitted on : Tuesday, October 10, 2006 - 12:55:10 PM
Last modification on : Thursday, February 7, 2019 - 2:36:01 PM
Long-term archiving on : Tuesday, April 6, 2010 - 7:09:58 PM

File

cr112021373080.pdf

Identifiers

  • HAL Id : inria-00105112, version 1

Collections

IWFHR10 | UNIV-RENNES1 | INSA-RENNES | IRISA-D6 | UNIV-RENNES | INSA-GROUPE

Citation

Solen Quiniou, Eric Anquetil. A Priori and A Posteriori Integration and Combination of Language Models in an On-line Handwritten Sentence Recognition System. Tenth International Workshop on Frontiers in Handwriting Recognition, Université de Rennes 1, Oct 2006, La Baule (France), France. ⟨inria-00105112⟩

Export

BibTeX TEI DC DCterms EndNote

Share

Metrics

Record views

136

Files downloads

156

Contact


archive-ouverte@inria.fr