Étude préliminaire de reconnaissance d'écriture sur des documents historiques

Adeline Granet 1, 2, 3 Emmanuel Morin 1, 2 Harold Mouchère 1, 3 Solen Quiniou 1, 2 Christian Viard-Gaudin 1, 3
2 TALN - Traitement Automatique du Langage Naturel
LS2N - Laboratoire des Sciences du Numérique de Nantes
3 IPI - Image Perception Interaction
LS2N - Laboratoire des Sciences du Numérique de Nantes
Abstract : ABSTRACT. This work cares about information retrieval in accounting registers of Italian comedy of the 18 th century. These documents contain precious information for human and social science researchers interested in the integration of the Italian actors during this century. Information retrieval in old documents which have never been studied before, is a long and difficult process. Each step asks an expertise : detection and segmentation into blocs, lines or words; extraction efficient features; and handwriting recognition. The BLSTM recurrent neural network with CTC decoding is the most popular solution which outperforms others for alignment between a transcription and an input sequence. This paper explains a preliminary investigation using this kind of recurrent neural network for the following task : identify the play's titles in multilingual historical documents using closed vocabulary that mainly contains named entities.
Complete list of metadatas

Cited literature [15 references]  Display  Hide  Download

https://hal.archives-ouvertes.fr/hal-01758573
Contributor : Adeline Granet <>
Submitted on : Wednesday, April 4, 2018 - 3:46:52 PM
Last modification on : Tuesday, March 26, 2019 - 9:25:22 AM

File

RJCRI2017.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : hal-01758573, version 1

Citation

Adeline Granet, Emmanuel Morin, Harold Mouchère, Solen Quiniou, Christian Viard-Gaudin. Étude préliminaire de reconnaissance d'écriture sur des documents historiques. Rencontre des Jeunes Chercheurs en Recherche d'Information (RJCRI), Mar 2017, Marseille, France. ⟨hal-01758573⟩

Share

Metrics

Record views

198

Files downloads

85