Développement d'une plateforme de crowdsourcing pour la transcription et la labellisation de documents anciens

Details

Supervisors

Deville, Yves

Faculty

Ecole polytechnique de Louvain

Degree label

Master [120] en sciences informatiques, à finalité spécialisée

Abstract

enTranscribing ancient documents is an essential task for historians, linguists and researchers interested in preserving and understanding the history of mankind. Today, with millions of scanned documents, researchers encounter difficulties in finding precise information and analysing the various documents in a transversal manner. The aim of this thesis is to develop a platform, named Archio, designed to facilitate the transcription and labelling of ancient documents. Thanks to crowdsourcing, this platform aims to speed up and improve the processing of historical texts. Crowdsourcing is a form of outsourcing in which a task is delegated to any group of people. This platform also aims to facilitate access to ancient knowledge for academic research and public education. The objective is also to provide useful data for training an artificial intelligence model to transcribe ancient documents. By adopting a global and generalist approach, this platform aims to improve the transcription of any type of document, regardless of subject, language or page layout.