News

TEKLIA signs a letter of intent with the Egyptian Ministry of Communications and Information Technology to implement AI-Powered projects

September 13, 2020

Paris - September 13th, 2020 TEKLIA is very happy to announce a collaboration with the Ministry of Communications and Information Technology (MCIT) in Egypt to develop AI-based technology to automatically understand handwritten historical documents in Arabic. TEKLIA has already demonstrated to MCIT a system for automatic transcription of modern Arabic and the collaboration will allow to develop a platform for semantic information extraction able to cope with the challenges of handwritten and printed historical documents in Arabic. Read more

Arkindex in practice: import IIIF images

April 15, 2020

This article will guide you through the steps necessary to import images from a IIIF server to Arkindex. You will learn how to communicate with the Arkindex Server via the API. You will also get a glimpse of how documents are organized in Arkindex. In the following article, we will assume you have a list of urls of images already uploaded on an IIIF server, that you want to import to Arkindex. Read more

Internship in Deep Learning models for historical handwriting recognition

November 16, 2019

As for many applications of computer vision, significant progress have been achieved recently in the field of handwriting recognition, thanks to Deep Learning. Handwriting recognition systems are very similar to speech recognition ones: convolutional neural networks are used to extract features, recurrent neural networks model the sequences of characters and attention mechanisms are used to follow the writing and decode the sequence of words. Similarly to other applications of deep learning, the performance are strongly correlated to the quantity of good quality annotated data. Read more

Offre de stage/emploi en Développement Back/Front et DevOps

November 16, 2019

Dans le cadre du développement de la plateforme Arkindex, TEKLIA cherche à renforcer son équipe de développement par le recrutement d’un stagiaire en pré-embauche CDI ou d’un développeur en CDI. L’objectif est de participer à la mise en place d’un système complet de compréhension automatique de documents numérisés par des techniques de machine learning et deep learning. Cette plateforme est développée en collaboration avec des historiens paléographes du CNRS-IRHT (https://www.irht.cnrs.fr/), des informaticiens linguistes du LS2N (https://ls2n. Read more

TEKLIA will assist the Paris School of Economics in studying administrative documents in Nepali

October 2, 2019

TEKLIA is very happy to assist the Paris School of Economics in a research project dedicated to the study of local governance of forest resources in Nepal, GoLFor-DEEPN. This project, funded by the French Agency for Research, will provide the first evaluation at the national level of the effect of handing over the management of forest from the Department of Forest to Community Forest User Groups. This analysis will be partly based on automatic information extraction from administrative records in Nepali. Read more

TEKLIA presents two research papers at ICDAR2019

September 15, 2019

TEKLIA will present two research papers at the 15th International Conference on Document Analysis and Recognition, organised by University of Technology Sydney (UTS), Australia and hold at the International Convention Centre (ICC) Sydney. A first paper, introduces a new data set of fully annotated book of hours and first results on the automatic analysis of their structure. It will be presented at the 5th International Workshop on Historical Document Imaging and Processing : Read more

TEKLIA's CaptchAN project awarded a grant from the Ministry of Culture

June 13, 2019

TEKLIA and their partners Decalog, the French Archives Nationales, Bibliothèque Nationale de France and the Musée de Bretagne are very proud to announce that their project CaptchAN has been awarded an Innovative Digital Service grant from the French Ministry of Culture. The CaptchAN project will develop a captcha service using cultural and patrimonial data provided by partner cultural institutions. The goal is twofold : to provide an alternative to the commercial captcha services that make use of public websites users’ free labour to improve private data, and to accelerate the indexation of public archives and libraries’ documents. Read more

TEKLIA joins the BALSAC project

April 12, 2019

TEKLIA are very happy to announce that they have joined the BALSAC project, for which they will automate the handwriting recognition and information extraction. The BALSAC project maintains and updates a database of digitized vital event records (birth, death and marriage certificates) from all over Quebec, from the first European settlements in the 17th century up to the present day. These records are interconnected using a linkage method based on nominative information, which allows for the automatic recreation of genealogical relations and kinship structures in the Quebec population. Read more

Emanuela Boros Thesis Defense

September 26, 2018

Emanuela Boros, will defend her PhD Thesis entitled neural methods for event extraction on September 27, 2018 at 2pm. The defence will take place in conférence room, bâtiment 507, LIMSI, Orsay. Abstract With the increasing amount of data and the exploding number data sources, the extraction of information about events, whether from the perspective of acquiring knowledge or from a more directly operational perspective, becomes a more and more obvious need. Read more